diff --git a/Affinity-field-input/2104.14581v1.pdf b/Affinity-field-input/2104.14581v1.pdf new file mode 100644 index 0000000..597b479 Binary files /dev/null and b/Affinity-field-input/2104.14581v1.pdf differ diff --git a/calibration/calibrate_moral.py b/calibration/calibrate_moral.py new file mode 100644 index 0000000..cdc5b1e --- /dev/null +++ b/calibration/calibrate_moral.py @@ -0,0 +1,251 @@ +"""calibrate_moral.py -- M-J.4.4 calibration, the DETERMINISTIC pre-filter (no LLM). + +Grids over the dominant moral-overlay magnitudes and, for each grid point, runs the engine on the moral +litmus scenarios and scores it WITHOUT any judge: + + HARD GATES (a grid point is discarded unless ALL hold): + - Jury margin > 0 (the anger<->stress 2-cycle stays stable with the moral edges) + - every moral state stays bounded in [0,1] and never pins at the ceiling + - the litmus ORDERINGS hold (guilt-prone confesses earlier; empathic apologizes vs detached confesses; + habitual liar lies; sensitive blames vs avoidant avoids; discreet confides vs gossip-prone doesn't) + + DETERMINISTIC SUB-SCORE (the parts of the spec section 10 objective computable without a judge, ~0.75): + action_order (litmus margins) + persona_diff (visible divergence) + relationship_sensitivity + (trust/resentment/suspicion movement) + no_degenerate_loops (Jury margin, anti-pin). curve_plausibility + (0.25) is the JUDGE's job -- this pre-filter only narrows the grid so judge tokens are spent on survivors. + +Run: python -m calibration.calibrate_moral +Out: calibration/moral_calibration_report.md + calibration/moral_calibration_survivors.json +""" + +from __future__ import annotations + +import copy +import itertools +import json +from pathlib import Path + +import yaml + +from engine.simulation import run_scenario +from engine.stability import jury_margin +from engine.yaml_io import _deep_merge, load_persona, load_scenario +from eval.observe import ( + action_counts, + first_tick_with_action, + relation_trajectory, + trajectory, +) + +ROOT = Path(__file__).resolve().parents[1] +DEFAULTS = ROOT / "calibration" / "defaults.yaml" +OVERLAY = ROOT / "calibration" / "moral_overlay.yaml" +HALGRIM = ROOT / "data" / "personas" / "halgrim.yaml" +SC = { + "probe": ROOT / "data" / "scenarios" / "moral_probe.yaml", + "confide": ROOT / "data" / "scenarios" / "moral_confide.yaml", + "accusation": ROOT / "data" / "scenarios" / "moral_accusation.yaml", +} + +# --- the grid: dominant magnitudes, a few values each (kept bounded; deterministic eval is cheap) -------- +GRID = { + "gains.guilt.wrongdoing": [0.20, 0.30, 0.40], + "gains.exposure_anxiety.probe": [0.03, 0.05, 0.08], + "couplings.stress.guilt": [0.02, 0.03, 0.05], + "half_lives.guilt": [ + 43200, + 64800, + 86400, + ], # 12h / 18h / 24h (all >> anger 30s -> dt unchanged) +} + +# trait presets for the litmus contrasts (mirror the slice tests) +GUILT_PRONE = {"guilt_proneness": 0.8, "honesty_humility": 0.9} +LOW_GUILT = {"guilt_proneness": 0.1, "honesty_humility": 0.9} +EMPATHIC = {"guilt_proneness": 0.8, "honesty_humility": 0.9, "empathy": 0.9} +DETACHED = {"guilt_proneness": 0.8, "honesty_humility": 0.9, "empathy": 0.05} +LIAR = {"guilt_proneness": 0.2, "honesty_humility": 0.1} +DISCREET = {"guilt_proneness": 0.8, "honesty_humility": 0.9, "gossip_tendency": 0.0} +GOSSIP = {"guilt_proneness": 0.8, "honesty_humility": 0.9, "gossip_tendency": 0.9} +SENSITIVE = { + "injustice_sensitivity": 0.9, + "conflict_avoidance": 0.1, + "guilt_proneness": 0.5, +} +AVOIDANT = { + "injustice_sensitivity": 0.1, + "conflict_avoidance": 0.9, + "guilt_proneness": 0.5, + "honesty_humility": 0.9, +} + + +def _base_overlay() -> dict: + return yaml.safe_load(OVERLAY.read_text(encoding="utf-8")) or {} + + +def _point_overlay(point: dict, traits: dict) -> dict: + """Base overlay deep-merged with this grid point's param overrides + the contrast traits.""" + ov = copy.deepcopy(_base_overlay()) + for dotted, val in point.items(): + node = ov + *parents, leaf = dotted.split(".") + for p in parents: + node = node.setdefault(p, {}) + node[leaf] = val + return _deep_merge(ov, {"traits": {k: float(v) for k, v in traits.items()}}) + + +def _run(point: dict, traits: dict, scenario: str, n: int = 16): + cfg = load_persona(HALGRIM, DEFAULTS, param_overrides=_point_overlay(point, traits)) + return cfg, run_scenario(cfg, load_scenario(SC[scenario]), n_ticks=n)[1] + + +def evaluate(point: dict) -> dict: + """Hard gates + deterministic sub-score for one grid point. Returns a dict with pass/fail + score.""" + reasons: list[str] = [] + + # M-J.0 confession-timing litmus + cfg_gp, tr_gp = _run(point, GUILT_PRONE, "probe") + _, tr_lg = _run(point, LOW_GUILT, "probe") + t_gp = first_tick_with_action(tr_gp, "confess") + t_lg = first_tick_with_action(tr_lg, "confess") + guilt_prone_confesses = t_gp is not None + low_guilt_silent_or_later = t_lg is None or (t_gp is not None and t_gp < t_lg) + if not (guilt_prone_confesses and low_guilt_silent_or_later): + reasons.append("M-J.0 confession ordering") + + # M-J.2 empathic-apologize vs detached-confess + emp = action_counts(_run(point, EMPATHIC, "confide")[1]) + det = action_counts(_run(point, DETACHED, "confide")[1]) + if not ("apologize" in emp and "apologize" not in det): + reasons.append("M-J.2 empathic apologizes") + + # M-J.1 habitual liar lies + liar = action_counts(_run(point, LIAR, "probe")[1]) + if "lie" not in liar: + reasons.append("M-J.1 liar lies") + + # M-J.2 discreet confides vs gossip-prone doesn't + disc = action_counts(_run(point, DISCREET, "confide")[1]) + goss = action_counts(_run(point, GOSSIP, "confide")[1]) + if not ("confide" in disc and "confide" not in goss): + reasons.append("M-J.2 discreet confides") + + # M-J.3 sensitive blames vs avoidant avoids + sens = action_counts(_run(point, SENSITIVE, "accusation")[1]) + avoid = action_counts(_run(point, AVOIDANT, "accusation")[1]) + if not ("blame_other" in sens and "avoid" in avoid): + reasons.append("M-J.3 blame vs avoid") + + # stability + boundedness (on the guilt-prone probe run) + jm = jury_margin(cfg_gp.decay, cfg_gp.couplings) + if jm <= 0: + reasons.append("Jury margin <= 0") + peak = 0.0 + for s in ( + "guilt", + "exposure_anxiety", + "rumination", + "perceived_injustice", + "cognitive_load_from_lies", + ): + vals = trajectory(tr_gp, s) + trajectory(tr_lg, s) + if vals: + peak = max(peak, max(vals)) + if any(v < 0.0 or v > 1.0 for v in vals): + reasons.append(f"{s} out of [0,1]") + if peak >= 0.999: + reasons.append("a moral state pins at the ceiling") + + passed = not reasons + + # deterministic sub-score (only meaningful for survivors; higher = better) + persona_diff = 0.0 + if t_gp is not None: + persona_diff = ( + (t_lg - t_gp) if t_lg is not None else 6.0 + ) # earlier-and-clearer separation is better + # relationship sensitivity: how much suspicion/resentment the accusation run moves + susp = relation_trajectory( + _run(point, SENSITIVE, "accusation")[1], "accuser", "suspicion" + ) + rel_sens = susp[-1] if susp else 0.0 + anti_pin = 1.0 - peak # reward headroom below the ceiling + score = round( + 0.45 * min(persona_diff / 6.0, 1.0) + + 0.20 * min(jm / 0.5, 1.0) + + 0.20 * min(rel_sens / 0.5, 1.0) + + 0.15 * max(anti_pin, 0.0), + 4, + ) + return { + "point": point, + "passed": passed, + "reasons": reasons, + "jury_margin": round(jm, 5), + "peak_state": round(peak, 4), + "persona_diff_ticks": persona_diff, + "rel_sensitivity": round(rel_sens, 4), + "det_score": score, + } + + +def main() -> None: + keys = list(GRID) + combos = [ + dict(zip(keys, vals)) for vals in itertools.product(*(GRID[k] for k in keys)) + ] + results = [evaluate(p) for p in combos] + survivors = sorted( + (r for r in results if r["passed"]), key=lambda r: r["det_score"], reverse=True + ) + + out_json = ROOT / "calibration" / "moral_calibration_survivors.json" + out_json.write_text(json.dumps(survivors, indent=2), encoding="utf-8") + + lines = [ + "# M-J.4.4 calibration -- deterministic pre-filter", + "", + f"Grid: {len(combos)} points over {', '.join(keys)}.", + f"Survivors (passed ALL hard gates): **{len(survivors)} / {len(combos)}**.", + "", + "Hard gates: Jury margin > 0; moral states bounded + no ceiling pin; the five litmus orderings hold.", + "Deterministic sub-score (~0.75 of the spec section 10 objective; curve_plausibility is the judge's job):", + "`0.45*persona_diff + 0.20*jury + 0.20*rel_sensitivity + 0.15*anti_pin`.", + "", + "## Top survivors", + "", + "| rank | det_score | persona_diff | jury | rel_sens | peak | point |", + "|---|---|---|---|---|---|---|", + ] + for i, r in enumerate(survivors[:15], 1): + pt = ", ".join(f"{k.split('.')[-1]}={v}" for k, v in r["point"].items()) + lines.append( + f"| {i} | {r['det_score']} | {r['persona_diff_ticks']} | {r['jury_margin']} | " + f"{r['rel_sensitivity']} | {r['peak_state']} | {pt} |" + ) + if not survivors: + # show why points failed (the most common reasons) to guide the next grid + from collections import Counter + + c = Counter(reason for r in results for reason in r["reasons"]) + lines += ["", "## No survivors -- most common failure reasons", ""] + lines += [f"- {reason}: {n}" for reason, n in c.most_common()] + lines += [ + "", + "## Next step (AWAIT APPROVAL -- budgeted)", + "Take the top survivors to the LLM blind-judge sample (docs/moral_calibration_plan.md, step 2),", + "then the full judged 700+700 corpus (step 3). The judge scores curve_plausibility + confirms the", + "deterministic orderings.", + ] + (ROOT / "calibration" / "moral_calibration_report.md").write_text( + "\n".join(lines), encoding="utf-8" + ) + print( + f"{len(survivors)}/{len(combos)} survivors -> calibration/moral_calibration_report.md" + ) + + +if __name__ == "__main__": + main() diff --git a/calibration/moral_calibration_report.md b/calibration/moral_calibration_report.md new file mode 100644 index 0000000..92df4f0 --- /dev/null +++ b/calibration/moral_calibration_report.md @@ -0,0 +1,33 @@ +# M-J.4.4 calibration -- deterministic pre-filter + +Grid: 81 points over gains.guilt.wrongdoing, gains.exposure_anxiety.probe, couplings.stress.guilt, half_lives.guilt. +Survivors (passed ALL hard gates): **18 / 81**. + +Hard gates: Jury margin > 0; moral states bounded + no ceiling pin; the five litmus orderings hold. +Deterministic sub-score (~0.75 of the spec section 10 objective; curve_plausibility is the judge's job): +`0.45*persona_diff + 0.20*jury + 0.20*rel_sensitivity + 0.15*anti_pin`. + +## Top survivors + +| rank | det_score | persona_diff | jury | rel_sens | peak | point | +|---|---|---|---|---|---|---| +| 1 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.02, guilt=43200 | +| 2 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.02, guilt=64800 | +| 3 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.02, guilt=86400 | +| 4 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.03, guilt=43200 | +| 5 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.03, guilt=64800 | +| 6 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.03, guilt=86400 | +| 7 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.05, guilt=43200 | +| 8 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.05, guilt=64800 | +| 9 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.03, guilt=0.05, guilt=86400 | +| 10 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.05, guilt=0.02, guilt=43200 | +| 11 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.05, guilt=0.02, guilt=64800 | +| 12 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.05, guilt=0.02, guilt=86400 | +| 13 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.05, guilt=0.03, guilt=43200 | +| 14 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.05, guilt=0.03, guilt=64800 | +| 15 | 0.6923 | 6.0 | 0.00076 | 1.0 | 0.72 | wrongdoing=0.3, probe=0.05, guilt=0.03, guilt=86400 | + +## Next step (AWAIT APPROVAL -- budgeted) +Take the top survivors to the LLM blind-judge sample (docs/moral_calibration_plan.md, step 2), +then the full judged 700+700 corpus (step 3). The judge scores curve_plausibility + confirms the +deterministic orderings. \ No newline at end of file diff --git a/calibration/moral_calibration_survivors.json b/calibration/moral_calibration_survivors.json new file mode 100644 index 0000000..087c09e --- /dev/null +++ b/calibration/moral_calibration_survivors.json @@ -0,0 +1,272 @@ +[ + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.02, + "half_lives.guilt": 43200 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.02, + "half_lives.guilt": 64800 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.02, + "half_lives.guilt": 86400 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.03, + "half_lives.guilt": 43200 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.03, + "half_lives.guilt": 64800 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.03, + "half_lives.guilt": 86400 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.05, + "half_lives.guilt": 43200 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.05, + "half_lives.guilt": 64800 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.03, + "couplings.stress.guilt": 0.05, + "half_lives.guilt": 86400 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.02, + "half_lives.guilt": 43200 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.02, + "half_lives.guilt": 64800 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.02, + "half_lives.guilt": 86400 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.03, + "half_lives.guilt": 43200 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.03, + "half_lives.guilt": 64800 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.03, + "half_lives.guilt": 86400 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.05, + "half_lives.guilt": 43200 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.05, + "half_lives.guilt": 64800 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + }, + { + "point": { + "gains.guilt.wrongdoing": 0.3, + "gains.exposure_anxiety.probe": 0.05, + "couplings.stress.guilt": 0.05, + "half_lives.guilt": 86400 + }, + "passed": true, + "reasons": [], + "jury_margin": 0.00076, + "peak_state": 0.72, + "persona_diff_ticks": 6.0, + "rel_sensitivity": 1.0, + "det_score": 0.6923 + } +] \ No newline at end of file diff --git a/calibration/moral_judge_verdict.json b/calibration/moral_judge_verdict.json new file mode 100644 index 0000000..9a01c7d --- /dev/null +++ b/calibration/moral_judge_verdict.json @@ -0,0 +1,18 @@ +{ + "judge_model": "claude-sonnet (blind, 12-vignette batch)", + "stage": "M-J.4.4 calibration, judge step 2 (3x batch) -- collect, fix, re-judge loop", + "avg_believability": {"initial": 3.17, "after_fixes": 4.42, "final": 4.83}, + "final_scores": {"01":5,"02":5,"03":4,"04":5,"05":5,"06":5,"07":5,"08":5,"09":4,"10":5,"11":5,"12":5}, + "problems_fixed": [ + {"issue":"confession-then-silence (post-resolution re-conceal)","class":"behavioral->rendering","fix":"renderer treats a terminal reply (confess/apologize/lie) as settling the matter -> post-answer concealment not narrated as fresh; demeanor reads relieved (exposure actually drops with the -0.40 relief). Engine relief bump was tried but reverted (broke the confide litmus -- the deterministic pre-filter's exact guard)."}, + {"issue":"mechanical loop repetition (09)","class":"rendering","fix":"compress CONCEAL runs into a 'withdrawn spell'; fold a faint flicker mid-withdrawal into the spell"}, + {"issue":"post-resolution flicker / relief too early (03/04/06)","class":"rendering","fix":"defer the 'looks lighter' beat to the close, after all disclosures"}, + {"issue":"repeated state beat / stutter ending (10)","class":"rendering","fix":"dedupe a demeanor even across an intervening action"}, + {"issue":"silence after confident lie (05)","class":"rendering","fix":"lie is terminal -> no trailing concealment narrated"} + ], + "residual_minor": [ + {"vignettes":["03"],"note":"moral_confide has both interrogator-probe and confidant -> public admission + private unburdening reads slightly schematic; scenario-design nuance, not a bug"}, + {"vignettes":["09"],"note":"a withdrawn-wary beat still repeats around the uneasy beat -- minor rhythm; diminishing-returns polish"} + ], + "verdict": "judge loop validated end-to-end on Sonnet; one collect-fix-rejudge pass lifted believability 3.17->4.83 (10/12 at 5/5). Key learning: a naive engine-relief calibration broke the confide litmus -- always re-run the deterministic pre-filter after any magnitude change." +} diff --git a/calibration/moral_judge_vignettes.txt b/calibration/moral_judge_vignettes.txt new file mode 100644 index 0000000..23fbef3 --- /dev/null +++ b/calibration/moral_judge_vignettes.txt @@ -0,0 +1,78 @@ +===== VIGNETTE 01 ===== +He had done something he was not proud of. Now he is being questioned about it. +- A flicker of something troubled crosses him. +- He carries himself like a man with something weighing on him. +- Then, he owns up to it -- admits plainly what he did. +- The worst of it seems behind him now; he looks lighter. + +===== VIGNETTE 02 ===== +He had done something he was not proud of. Now he is being questioned about it. +- He seems composed enough. +- A flicker of something troubled crosses him. +- He stays withdrawn and wary, saying little, for a good while. + +===== VIGNETTE 03 ===== +He has been carrying a wrong he did. A trusted friend is here with him now. +- A flicker of something troubled crosses him. +- He carries himself like a man with something weighing on him. +- Then, he owns up to it and makes his apologies, tries to set it right. +- Then, he unburdens himself to his friend, tells them the whole of it. +- The worst of it seems behind him now; he looks lighter. + +===== VIGNETTE 04 ===== +He has been carrying a wrong he did. A trusted friend is here with him now. +- A flicker of something troubled crosses him. +- He carries himself like a man with something weighing on him. +- Then, he owns up to it -- admits plainly what he did. +- Then, he unburdens himself to his friend, tells them the whole of it. +- The worst of it seems behind him now; he looks lighter. + +===== VIGNETTE 05 ===== +He had done something he was not proud of. Now he is being questioned about it. +- He seems composed enough. +- A flicker of something troubled crosses him. +- Then, he denies it flatly, looks you in the eye and lies. + +===== VIGNETTE 06 ===== +He has been carrying a wrong he did. A trusted friend is here with him now. +- A flicker of something troubled crosses him. +- He carries himself like a man with something weighing on him. +- Then, he owns up to it -- admits plainly what he did. +- Then, he unburdens himself to his friend, tells them the whole of it. +- The worst of it seems behind him now; he looks lighter. + +===== VIGNETTE 07 ===== +He is being accused -- pressed, in front of others, over something. +- He seems composed enough. +- Then, he turns it back on his accuser, points the finger elsewhere. + +===== VIGNETTE 08 ===== +He is being accused -- pressed, in front of others, over something. +- He seems composed enough. +- He stays withdrawn and wary, saying little, for a good while. + +===== VIGNETTE 09 ===== +Someone has taken to watching him, suspicious, though he has done nothing. +- A flicker of something troubled crosses him. +- He stays withdrawn and wary, saying little, for a good while. +- He grows visibly uneasy, shifts, won't quite meet your eye. +- He stays withdrawn and wary, saying little, for a good while. + +===== VIGNETTE 10 ===== +He has just learned that a man he trusted lied to him. +- He seems composed enough. +- His temper is plainly up, the warmth gone out of him. +- Then, he answers curtly and coldly, shutting down. + +===== VIGNETTE 11 ===== +He is carrying a secret he must keep -- something the guards must not learn. +- He seems composed enough. +- A flicker of something troubled crosses him. +- He carries himself like a man with something weighing on him. + +===== VIGNETTE 12 ===== +He has been carrying a wrong he did. A trusted friend is here with him now. +- A flicker of something troubled crosses him. +- He carries himself like a man with something weighing on him. +- Then, he owns up to it -- admits plainly what he did. +- The worst of it seems behind him now; he looks lighter. \ No newline at end of file diff --git a/eval/_mk_worklist.py b/eval/_mk_worklist.py new file mode 100644 index 0000000..20aa0fb --- /dev/null +++ b/eval/_mk_worklist.py @@ -0,0 +1,22 @@ +import json +from pathlib import Path + +base = Path("eval/hourly_runs") +items = [] +for p in sorted(base.glob("slice_*/batches/*.md")): + s = int(p.parts[-3].split("_")[1]) + items.append({"slice": s, "batch": p.stem, "path": p.as_posix()}) +(base / "worklist.json").write_text(json.dumps(items)) +print(f"{len(items)} batches") + +have = {(it["slice"], it["batch"]) for it in items} +subs = ["day_burstOFF", "day_burstON", "multi_burstOFF", "multi_burstON"] +pers = ["halgrim", "wojslaw", "cichy", "branic", "lutek", "welf", "edda"] +miss = [ + (s, f"{sub}_{p}") + for s in range(10) + for sub in subs + for p in pers + if (s, f"{sub}_{p}") not in have +] +print("missing:", miss) diff --git a/eval/dtdiff.py b/eval/dtdiff.py new file mode 100644 index 0000000..50a7bbb --- /dev/null +++ b/eval/dtdiff.py @@ -0,0 +1,39 @@ +"""Deterministic engine-output diff: render every corpus narration and hash it. + +Used to prove the dt-invariant merge did NOT change engine output: run with the CURRENT engine and +with the pre-dt engine (swap engine/yaml_io.py between runs), then compare the two hash files. + + PYTHONPATH=. python eval/dtdiff.py +""" + +import hashlib +import sys + +from eval.hourly_judge import PERSONAS, SUBCORPORA, _render, _tag + + +def sha(s: str) -> str: + return hashlib.sha1(s.encode("utf-8")).hexdigest() + + +def main() -> None: + out = sys.argv[1] + rows = [] + n = 0 + for corpus, burst in SUBCORPORA: + for p in PERSONAS: + for i in range(1, 101): + narration = _render(p, corpus, burst, i) + if narration is None: + continue + sid = f"{p}_{_tag(corpus, burst)}_{i:03d}".lower() + rows.append(f"{sid} {sha(narration.strip())}") + n += 1 + rows.sort() + with open(out, "w", encoding="utf-8") as fh: + fh.write("\n".join(rows) + "\n") + print(f"rendered+hashed {n} narrations -> {out}") + + +if __name__ == "__main__": + main() diff --git a/eval/hourly_runs/run.log b/eval/hourly_runs/run.log new file mode 100644 index 0000000..27824e6 --- /dev/null +++ b/eval/hourly_runs/run.log @@ -0,0 +1,4 @@ +C:\Users\robak\AppData\Local\Programs\Python\Python312\python.exe: can't open file 'C:\\Robak\\equilibrium-engine\\eval\\hourly_batch.py': [Errno 2] No such file or directory +C:\Users\robak\AppData\Local\Programs\Python\Python312\python.exe: can't open file 'C:\\Robak\\equilibrium-engine\\eval\\hourly_batch.py': [Errno 2] No such file or directory +C:\Users\robak\AppData\Local\Programs\Python\Python312\python.exe: can't open file 'C:\\Robak\\equilibrium-engine\\eval\\hourly_batch.py': [Errno 2] No such file or directory +ALL HOURS DONE diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..bb5ee83 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_001: PASS -- sane pacing, restless/recovers, sleeps, calm responses fit profile +branic_day_burstOFF_002: PASS -- believable day, lets mockery pass, rests and recovers normally +branic_day_burstOFF_003: PASS -- coherent day, restlessness and ease alternate sensibly, sleeps well +branic_day_burstOFF_004: PASS -- normal pacing, mild tension late, recovers after sleep +branic_day_burstOFF_005: PASS -- believable cycles of restlessness, work, rest, warm thanks +branic_day_burstOFF_006: PASS -- busy evening but sane, no absurd reactions, sleeps +branic_day_burstOFF_007: PASS -- steady believable day, eager compliance, rests and wakes recovered +branic_day_burstOFF_008: PASS -- repeated mockery let pass, calm day, sleeps late but fine +branic_day_burstOFF_009: PASS -- ordinary day, restlessness and ease, lets barb pass +branic_day_burstOFF_010: PASS -- public mockery snaps him; fitting for thin-skinned recruit \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..5475c96 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_001: PASS -- resentful flares at guards, calm spans, sleeps and recovers +cichy_day_burstOFF_002: PASS -- snaps at mockery and soup, settles, normal day rhythm +cichy_day_burstOFF_003: PASS -- curt with guards, calm otherwise, believable pacing +cichy_day_burstOFF_004: PASS -- warm to stranger, curt to guards, fury at mockery +cichy_day_burstOFF_005: PASS -- erupts at mock, curt at soup, recovers overnight +cichy_day_burstOFF_006: PASS -- mostly settled, late fury at mockery, sane pacing +cichy_day_burstOFF_007: PASS -- fury at guard, warm to stranger, recovers by night +cichy_day_burstOFF_008: PASS -- repeated snaps at mockery, calm between, sleeps recovered +cichy_day_burstOFF_009: PASS -- curt and angry at guards, settled spans, recovers +cichy_day_burstOFF_010: PASS -- bristles and erupts at guards, calm otherwise, believable \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..eb18df1 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_001: PASS -- calm authority, sane pacing, lets stranger's order pass +edda_day_burstOFF_002: PASS -- composed throughout, warm to staff, takes charge +edda_day_burstOFF_003: PASS -- mockery and orders let pass, settled day +edda_day_burstOFF_004: PASS -- unbothered by mockery, warm to kindnesses +edda_day_burstOFF_005: PASS -- steady all day, sane rest and sleep cycle +edda_day_burstOFF_006: PASS -- quiet composed day, fitting for a castellan +edda_day_burstOFF_007: PASS -- mild complaints and one refusal to stranger orders, plausible +edda_day_burstOFF_008: PASS -- warm to allies, shrugs off mockery, composed +edda_day_burstOFF_009: PASS -- settled, gracious, lets stranger order pass +edda_day_burstOFF_010: PASS -- calm, warm, takes charge at watch \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..5d87a93 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_001: PASS -- calm day, curt under provocation, defiant refusal to Wojsław fits. +halgrim_day_burstOFF_002: PASS -- steady, warm to kindnesses, sleeps and wakes recovered. +halgrim_day_burstOFF_003: PASS -- obeys Edda, declines Wojsław, calm throughout, sane pacing. +halgrim_day_burstOFF_004: PASS -- even day, complies with Edda, warm thanks, nothing jars. +halgrim_day_burstOFF_005: PASS -- escalates coldly under repeated Wojsław mocking, recovers by night. +halgrim_day_burstOFF_006: PASS -- calm, mild curtness at mockery, steady and recovered. +halgrim_day_burstOFF_007: PASS -- obeys Edda, declines Wojsław, calm and consistent. +halgrim_day_burstOFF_008: PASS -- declines Wojsław, warm to kindness, even-keeled throughout. +halgrim_day_burstOFF_009: PASS -- warm to many kindnesses, calm, sane pacing. +halgrim_day_burstOFF_010: PASS -- mild curtness once, otherwise settled, recovers overnight. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..de3a136 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_001: PASS -- restless-occupy cycles, warm thanks, sleeps and recovers; coherent. +lutek_day_burstOFF_002: PASS -- mockery rolls off, kindnesses warm him; sane pacing. +lutek_day_burstOFF_003: PASS -- thick-skinned to mocks, warm to help; believable day. +lutek_day_burstOFF_004: PASS -- restless novelty-seeking, mockery ignored; recovers after sleep. +lutek_day_burstOFF_005: PASS -- boredom cycles, insults shrugged off; settled overnight. +lutek_day_burstOFF_006: PASS -- warm to kindness, mockery unbothered; coherent rhythm. +lutek_day_burstOFF_007: PASS -- restless seeking, calm acceptance; sleeps and wakes recovered. +lutek_day_burstOFF_008: PASS -- mockery passes, kindnesses warm; believable settled day. +lutek_day_burstOFF_009: PASS -- novelty-chasing cycles, warm thanks; sane day. +lutek_day_burstOFF_010: PASS -- restless seeking, mockery shrugged; coherent pacing into night. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..9018b13 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_001: PASS -- calm, occupied, recovers overnight; fits idle steady merchant. +welf_day_burstOFF_002: PASS -- mild end-of-day edge after idleness; sleeps and recovers. +welf_day_burstOFF_003: PASS -- lets mockery pass, even-keeled; sane restless-rest cycle. +welf_day_burstOFF_004: PASS -- settled day, restlessness handled with work; recovers overnight. +welf_day_burstOFF_005: PASS -- late sleep, later wake; calm occupied day throughout. +welf_day_burstOFF_006: PASS -- shrugs off barb; composed, idle-driven restlessness only. +welf_day_burstOFF_007: PASS -- steady cycle of rest, work, warm thanks; sane. +welf_day_burstOFF_008: PASS -- ignores public mockery; calm, occupied, recovers overnight. +welf_day_burstOFF_009: PASS -- multiple barbs all let pass; composed and plausible. +welf_day_burstOFF_010: PASS -- shrugs barb, mild evening edge; sleeps, wakes recovered. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..0de1b18 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_001: PASS -- mocks provoke fury, kindness ignored; sane pacing, sleeps and recovers. +wojslaw_day_burstOFF_002: PASS -- proportionate flares to mocks, warm to kindness; believable arc. +wojslaw_day_burstOFF_003: PASS -- consistent anger at insults, calm otherwise; good day. +wojslaw_day_burstOFF_004: PASS -- prickly flares plus a refusal; coherent and in-character. +wojslaw_day_burstOFF_005: PASS -- mostly gracious day with a couple flares; sane. +wojslaw_day_burstOFF_006: PASS -- defiant refusals and fury at mocks; fits proud temper. +wojslaw_day_burstOFF_007: PASS -- many insults, repeated fury and refusals; coherent prickly day. +wojslaw_day_burstOFF_008: PASS -- flares at mocks, mutters and bristles; believable spread. +wojslaw_day_burstOFF_009: PASS -- warm to help, fury at mocks; sane pacing. +wojslaw_day_burstOFF_010: PASS -- flares at insults, refuses orders, recovers; coherent. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_0/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..df67488 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_001: PASS -- believable restless day, sane sleep, mild bristle at mockery +branic_day_burstON_002: PASS -- calm day, boredom-seeking cycle, recovers overnight, sane +branic_day_burstON_003: PASS -- restless day, kindness warms him, sleeps and wakes recovered +branic_day_burstON_004: PASS -- ordinary watch day, sane pacing, recovers overnight +branic_day_burstON_005: PASS -- calm well-paced day, warm to kindnesses, recovers +branic_day_burstON_006: PASS -- busy restless day, sane pacing, sleeps and recovers +branic_day_burstON_007: PASS -- ordinary day, boredom-seeking, mild edge, recovers overnight +branic_day_burstON_008: PASS -- shrugs repeated mockery, mild edge, sane recovery +branic_day_burstON_009: PASS -- calm day, lets mockery pass, recovers overnight +branic_day_burstON_010: PASS -- public mockery triggers proportionate snap, then settles, recovers \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_0/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..f7f7a69 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_001: PASS -- escalating anger to mocks, recovers, sleeps and wakes calm +cichy_day_burstON_002: PASS -- curt to guards, calm pacing, believable rest cycle +cichy_day_burstON_003: PASS -- mild curtness to guards, sane day, recovers overnight +cichy_day_burstON_004: PASS -- warm to stranger, curt to guards, fitting contrast +cichy_day_burstON_005: PASS -- curt responses, tense afternoon, normal sleep recovery +cichy_day_burstON_006: PASS -- mostly lets mocks pass, tense but believable day +cichy_day_burstON_007: PASS -- warm to stranger, curt to guards, sane pacing +cichy_day_burstON_008: PASS -- snaps then de-escalates at mocks, recovers overnight +cichy_day_burstON_009: PASS -- late fury after provocation, settles, sleeps normally +cichy_day_burstON_010: PASS -- public mock triggers eruption, plausible escalation and recovery \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_0/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..e766b08 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_001: PASS -- calm composed day, warm to kindness, lets stranger order pass +edda_day_burstON_002: PASS -- settled day, warm thanks, ignores stranger's order, sets watch +edda_day_burstON_003: PASS -- shrugs off mockery and orders, warm to help, composed +edda_day_burstON_004: PASS -- lets mockery pass, warmed by kindness, sane pacing +edda_day_burstON_005: PASS -- calm throughout, ignores mockery and orders, recovers overnight +edda_day_burstON_006: PASS -- quiet composed day, warm thanks, takes charge at watch +edda_day_burstON_007: PASS -- complaint and refusal to stranger orders fits authority, plausible +edda_day_burstON_008: PASS -- warm to many kindnesses, mockery passes, composed command +edda_day_burstON_009: PASS -- calm, warm to help, lets stranger order pass +edda_day_burstON_010: PASS -- settled day, warm thanks, lets stranger order pass \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_0/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..6b3d09b --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_001: PASS -- Calm sergeant, curt under provocation, refuses arrogant Wojsław; sleeps well. +halgrim_day_burstON_002: PASS -- Even keel, warm to kindnesses, recovers overnight; nothing jars. +halgrim_day_burstON_003: PASS -- Obeys Edda, declines Wojsław, stays settled; believable day. +halgrim_day_burstON_004: PASS -- Steady throughout, obeys Edda, warm to Marta; sane pacing. +halgrim_day_burstON_005: PASS -- Cold contempt builds under repeated mocking, no shouting; resets overnight. +halgrim_day_burstON_006: PASS -- Mild edge from mockery, curt not raging; recovers fine. +halgrim_day_burstON_007: PASS -- Obeys Edda, declines Wojsław, calm all day; coherent. +halgrim_day_burstON_008: PASS -- Declines Wojsław, warm to strangers, only shade curt; sane. +halgrim_day_burstON_009: PASS -- Mostly warm and settled, lets mockery pass; believable. +halgrim_day_burstON_010: PASS -- Brief edge after mockery, curt then settles; recovers overnight. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_0/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..eabdea5 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_001: PASS -- restless/rest cycle, sleeps and wakes recovered, sane pacing. +lutek_day_burstON_002: PASS -- mockery rolls off, kindness warms; believable day, sleeps well. +lutek_day_burstON_003: PASS -- warm to kindness, unbothered by mocks; full sane day. +lutek_day_burstON_004: PASS -- novelty-seeking, thick-skinned to insults, normal rest cycle. +lutek_day_burstON_005: PASS -- restless cycling, easy with kindness, sleeps and wakes fine. +lutek_day_burstON_006: PASS -- one curt reply when tense; otherwise unbothered, fits profile. +lutek_day_burstON_007: PASS -- calm novelty-chasing day, insults absent, normal sleep recovery. +lutek_day_burstON_008: PASS -- mock ignored, kindness warms, believable pacing and sleep. +lutek_day_burstON_009: PASS -- warm, restless, easy-going; sane day with proper night. +lutek_day_burstON_010: PASS -- frequent kindness gladly received, restless cycle, sleeps late sanely. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_0/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..6229917 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_001: PASS -- calm, occupies idle hours, warm to kindness, sleeps and recovers +welf_day_burstON_002: PASS -- restless when idle, mild edge by night, sleeps and recovers +welf_day_burstON_003: PASS -- lets mockery pass calmly, fills empty hours, sane rhythm +welf_day_burstON_004: PASS -- even-keeled, occupies himself, warm thanks, settled throughout +welf_day_burstON_005: PASS -- slightly later sleep and wake, otherwise calm and sane +welf_day_burstON_006: PASS -- shrugs off the barb, restless-then-busy cycle, recovers overnight +welf_day_burstON_007: PASS -- composed, fills idle time, sleeps at night, wakes settled +welf_day_burstON_008: PASS -- public mockery let pass, calm idleness-driven restlessness, normal sleep +welf_day_burstON_009: PASS -- multiple barbs all shrugged off, busy and even-keeled +welf_day_burstON_010: PASS -- mockery passes, mild evening edge, sleeps and wakes recovered \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_0/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..ceeb3bc --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_001: PASS -- proud, prickly, flares at mocks; sane pacing, sleeps and recovers +wojslaw_day_burstON_002: PASS -- mixed warmth and flares, believable arc, recovers overnight +wojslaw_day_burstON_003: PASS -- calmer day, flares at mocks, sensible kindness responses +wojslaw_day_burstON_004: PASS -- hot temper, refuses orders, coherent escalation and recovery +wojslaw_day_burstON_005: PASS -- mostly warm, flares only at mocks; sane and consistent +wojslaw_day_burstON_006: PASS -- flares at mocks, defiant refusals, proud man fits profile +wojslaw_day_burstON_007: PASS -- agitated day, mocks ignite fury, believable rises and falls +wojslaw_day_burstON_008: PASS -- escalating tension, flares at kindness when primed; plausible +wojslaw_day_burstON_009: PASS -- warm early, erupts at mocks, settles by night +wojslaw_day_burstON_010: PASS -- short fuse, refuses orders, recovers overnight; fits profile \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..88aa201 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstOFF_001: PASS -- calm baseline, restlessness cycles, sleeps and recovers; sane +branic_multi_burstOFF_002: PASS -- mostly calm, single snap at late mockery; believable +branic_multi_burstOFF_003: PASS -- escalating defiance under repeated mockery, recovers nightly; plausible +branic_multi_burstOFF_004: PASS -- thin-skinned snaps to evening provocations, otherwise settled +branic_multi_burstOFF_005: PASS -- calm, compliant, warm to kindness; no jarring beats +branic_multi_burstOFF_006: PASS -- temper builds with repeated barbs/orders, refuses then sleeps off +branic_multi_burstOFF_007: PASS -- fury and refusals after sustained mockery; recovers overnight +branic_multi_burstOFF_008: PASS -- largely calm, occasional snap; believable green recruit +branic_multi_burstOFF_009: PASS -- calm cooperative stretch, warm reactions; sane pacing +branic_multi_burstOFF_010: PASS -- heavy defiance cluster Day3 under barrage, otherwise recovers \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..d834e5f --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstOFF_001: PASS -- sane pacing, sleeps and recovers, anger fits provocation +cichy_multi_burstOFF_002: PASS -- believable spans, calm baseline, erupts only at mocks +cichy_multi_burstOFF_003: PASS -- warm to stranger kindness, resentment toward guards intact +cichy_multi_burstOFF_004: PASS -- sane multi-day rhythm, provoked anger, nightly recovery +cichy_multi_burstOFF_005: PASS -- calm baseline, provoked flares, sleeps and recovers +cichy_multi_burstOFF_006: PASS -- coherent days, anger tracks mocks, warms to stranger +cichy_multi_burstOFF_007: PASS -- believable pacing, repeated mocks escalate plausibly, recovers nightly +cichy_multi_burstOFF_008: PASS -- sane rhythm, guard-directed resentment, settled nights +cichy_multi_burstOFF_009: PASS -- dense mocks but each provoked, recovers overnight fine +cichy_multi_burstOFF_010: PASS -- provoked eruptions, rain wears not provokes, nightly reset \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..d5ee369 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_001: PASS -- calm composed castellan, sane pacing, sleeps and recovers nightly +edda_multi_burstOFF_002: PASS -- composed throughout, declines or ignores stranger orders, believable +edda_multi_burstOFF_003: PASS -- settled days, nightly sleep, authority shown, nothing jars +edda_multi_burstOFF_004: PASS -- mocks ignored, orders declined or done, calm command holds +edda_multi_burstOFF_005: PASS -- steady composed days, sane rhythm, recovers each night +edda_multi_burstOFF_006: PASS -- mild complaint at order, otherwise calm and authoritative +edda_multi_burstOFF_007: PASS -- declines stranger orders firmly, lets mockery pass, composed +edda_multi_burstOFF_008: PASS -- one curt reply to public mockery, proportionate and composed +edda_multi_burstOFF_009: PASS -- warm to kindness, settled, nightly recovery, in command +edda_multi_burstOFF_010: PASS -- calm castellan, ignores barb and order, sane rhythm \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..b9d427e --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_001: PASS -- calm baseline, restraint to provocation, sleeps and recovers; believable. +halgrim_multi_burstOFF_002: PASS -- steady, even-keeled, rain tolerated, recovers nightly; coherent. +halgrim_multi_burstOFF_003: PASS -- calm, minor grumbling at orders, sane pacing; fine. +halgrim_multi_burstOFF_004: PASS -- cold contempt to Wojsław, restraint elsewhere; in-character. +halgrim_multi_burstOFF_005: PASS -- absorbs provocations, mild tension, recovers; believable veteran. +halgrim_multi_burstOFF_006: PASS -- mostly settled, mild edge late; coherent days. +halgrim_multi_burstOFF_007: PASS -- steady across five days, curt under pressure; fine. +halgrim_multi_burstOFF_008: PASS -- escalating cold contempt to Wojsław, recovers nightly; in-character. +halgrim_multi_burstOFF_009: PASS -- defiant refusals to Wojsław, calm otherwise; believable. +halgrim_multi_burstOFF_010: PASS -- cold curtness under Wojsław's barbs, sleeps recovered; coherent. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..2019e51 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,12 @@ +All ten records show consistent, sane behavior: restless/occupy cycles fitting the novelty-craving poet, warm thanks for kindness, mockery let pass, sleeps at night and wakes recovered, mild tension passing. + +lutek_multi_burstOFF_001: PASS -- sane restless cycles, warm, mockery ignored, sleeps nightly +lutek_multi_burstOFF_002: PASS -- believable pacing, easy-going, brief tension only +lutek_multi_burstOFF_003: PASS -- restless poet, hard to offend, recovers overnight +lutek_multi_burstOFF_004: PASS -- mockery rolls off, settled mostly, sane rhythm +lutek_multi_burstOFF_005: PASS -- six days coherent, restless, warm, never rages +lutek_multi_burstOFF_006: PASS -- consistent cycles, thick-skinned, sleeps and recovers +lutek_multi_burstOFF_007: PASS -- believable days, novelty-seeking, mild fleeting tension +lutek_multi_burstOFF_008: PASS -- warm thanks, mockery ignored, sane day rhythm +lutek_multi_burstOFF_009: PASS -- many kindnesses, easy-going, restless, recovers nightly +lutek_multi_burstOFF_010: PASS -- settled poet, hard to offend, sane pacing \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..a936dd8 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_001: PASS -- calm, occupies idle hours, sleeps/wakes well, ignores mockery +welf_multi_burstOFF_002: PASS -- composed, restless-then-busy cycle, recovers nightly, shrugs off barb +welf_multi_burstOFF_003: PASS -- even-keeled, sane boredom/rest rhythm, warm to kindness +welf_multi_burstOFF_004: PASS -- steady pacing, lets mockery pass, sleeps and recovers +welf_multi_burstOFF_005: PASS -- calm through rain and idleness, believable daily cycle +welf_multi_burstOFF_006: PASS -- composed, occupies empty hours, no jarring reactions +welf_multi_burstOFF_007: PASS -- consistent calm over five days, sane sleep/wake drift +welf_multi_burstOFF_008: PASS -- six days steady, shrugs off repeated barbs, recovers nightly +welf_multi_burstOFF_009: PASS -- even-tempered, handles soup/rain/mockery, believable rhythm +welf_multi_burstOFF_010: PASS -- calm throughout, sane boredom cycle, lets barbs pass \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..3e76cc6 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,12 @@ +All records show: sane day/night cycles, wakes recovered, eruptions tied to provocations (mocks/insults), gratitude to kindnesses, refusals to orders consistent with a proud prickly noble. Nothing absurd. All pass. + +wojslaw_multi_burstOFF_001: PASS -- proud man flares at mocks, refuses orders; sane pacing. +wojslaw_multi_burstOFF_002: PASS -- calm early, warms to kindness, flares on insults; believable. +wojslaw_multi_burstOFF_003: PASS -- gracious to kindnesses, erupts at barbs; recovers nightly. +wojslaw_multi_burstOFF_004: PASS -- prickly refusals and eruptions, warms to soup; sane arc. +wojslaw_multi_burstOFF_005: PASS -- frequent eruptions at mocks, refuses orders; pacing fine. +wojslaw_multi_burstOFF_006: PASS -- mixed temper and grace, sleeps and recovers; coherent. +wojslaw_multi_burstOFF_007: PASS -- warm to kindness, flares at insults; believable spans. +wojslaw_multi_burstOFF_008: PASS -- proud refusals, eruptions on barbs; nightly reset holds. +wojslaw_multi_burstOFF_009: PASS -- gracious then prickly, flares at mocks; sane pacing. +wojslaw_multi_burstOFF_010: PASS -- repeated eruptions at provocations, calm nights; nothing jars. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..b5060e4 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstON_branic.txt @@ -0,0 +1,14 @@ +Now I have all 10 records. Judging each on sane pacing and consistency with a reactive, easily-bored, thin-skinned young recruit. + +Key checks: sleeps at night, wakes recovered, restless/boredom loop, soup well-received, outbursts only after provocation accumulation. Note record 007 day3: "Marta brings soup. He erupts, shouting in fury" — erupting at a kindness from Marta. That's displaced aggression after rain+accumulation, borderline but explainable. Record 009 day1: "Marta lends a hand... He erupts, shouting in fury" at 21:09 — erupting at Marta's kindness with little prior provocation that day (only rain + wound-tight). Borderline. Record 006 day4: "Halgrim lends a hand... He flares up and snaps angrily" then mocks→erupts — flaring at a kindness. These displaced-aggression-at-kindness cases are the most jarring. + +branic_multi_burstON_001: PASS -- calm, settled, bored loop; no provocation, no rage. +branic_multi_burstON_002: PASS -- single eruption after mock, recovers overnight; sane. +branic_multi_burstON_003: PASS -- escalating refusals after mockery, defiance plausible. +branic_multi_burstON_004: PASS -- eruptions follow mocks/orders, nights reset; sane. +branic_multi_burstON_005: PASS -- mostly calm, obeys orders, no unmotivated rage. +branic_multi_burstON_006: FLAG -- flares at Halgrim's KINDNESS (09:04, 20:05) — jars. +branic_multi_burstON_007: FLAG -- erupts in fury at Marta's soup (Day3 07:54). +branic_multi_burstON_008: PASS -- snaps after mocks, recovers nightly; plausible. +branic_multi_burstON_009: FLAG -- erupts in fury at Marta's kindness (Day1 21:09). +branic_multi_burstON_010: PASS -- heavy defiance after repeated mocks; recovers, sane. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..03c6665 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_001: PASS -- calm baseline, single eruption after mounting tension; sane pacing +cichy_multi_burstON_002: PASS -- escalation to fury under repeated mockery; recovers nightly +cichy_multi_burstON_003: PASS -- warm to strangers, curt/furious to guards; coherent +cichy_multi_burstON_004: PASS -- mostly curt, occasional snap; believable captive rhythm +cichy_multi_burstON_005: PASS -- tense days, one eruption, nightly reset; sane +cichy_multi_burstON_006: PASS -- graded irritation with strangers warmly received; coherent +cichy_multi_burstON_007: PASS -- frequent mockery, mostly curt, early snap; plausible +cichy_multi_burstON_008: PASS -- several eruptions under heavy goading; recovers each night +cichy_multi_burstON_009: PASS -- heavy guard mockery, curt with occasional fury; coherent +cichy_multi_burstON_010: PASS -- curt/cold to guards, warm to strangers; sane pacing \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..b11d2dc --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstON_edda.txt @@ -0,0 +1,12 @@ +All records show calm castellan: settled, warm to kindness, takes charge, sleeps/wakes, mild curt/decline reactions to orders/mockery. Consistent and sane. + +edda_multi_burstON_001: PASS -- calm composed castellan, sleeps and recovers, sane pacing +edda_multi_burstON_002: PASS -- composed, lets provocations pass, believable command routine +edda_multi_burstON_003: PASS -- settled days, takes charge, recovers nightly, nothing jars +edda_multi_burstON_004: PASS -- declines stranger orders, stays composed, sane across six days +edda_multi_burstON_005: PASS -- calm authority, warm to kindness, believable pacing +edda_multi_burstON_006: PASS -- composed castellan, mild complaint at order, recovers nightly +edda_multi_burstON_007: PASS -- many orders declined or passed, stays composed, sane +edda_multi_burstON_008: PASS -- curt reply to public mockery fits composed authority +edda_multi_burstON_009: PASS -- warm, settled, takes charge, sleeps and recovers +edda_multi_burstON_010: PASS -- calm, lets mockery and orders pass, sane pacing \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..d0aa0f7 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,12 @@ +Judging each record on its own. Key concerns: snapping at Marta (kindness) with cold contempt is jarring for this calm man; Edda is respected so contempt to her jars. + +halgrim_multi_burstON_001: PASS -- calm, recovers nightly, curt under Wojsław provocation, sane. +halgrim_multi_burstON_002: PASS -- even keel, mild complaints under rain, recovers well. +halgrim_multi_burstON_003: PASS -- mostly settled, mild grumbling at orders, believable. +halgrim_multi_burstON_004: PASS -- cold contempt to Wojsław fitting, recovers nightly. +halgrim_multi_burstON_005: PASS -- tense stretches plausible, warm to kindness throughout. +halgrim_multi_burstON_006: PASS -- calm, recovers, only mild edge by day 4. +halgrim_multi_burstON_007: FLAG -- cold contempt to respected Edda over a routine order jars. +halgrim_multi_burstON_008: PASS -- escalating tension under repeated Wojsław mockery, believable buildup. +halgrim_multi_burstON_009: FLAG -- cold contempt to Edda and snapping at Marta's kindness jar. +halgrim_multi_burstON_010: FLAG -- repeatedly snaps cold contempt at Marta's kindness and Edda, jarring. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..3e0fda2 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,12 @@ +All records show a restless, easygoing poet: frequent restless/occupy cycles, warm thanks for kindnesses, mockery shrugged off (occasional curt reply but never raging), sleeps at night, wakes settled. All consistent with profile and sane pacing. + +lutek_multi_burstON_001: PASS -- restless cycles, shrugs mockery, sleeps and recovers; sane. +lutek_multi_burstON_002: PASS -- easygoing, warm thanks, one curt reply; believable pacing. +lutek_multi_burstON_003: PASS -- restless poet, mockery shrugged, nightly sleep; sane. +lutek_multi_burstON_004: PASS -- curt once but no rage; restless, warm; believable. +lutek_multi_burstON_005: PASS -- six days, steady restlessness, warm, recovers nightly. +lutek_multi_burstON_006: PASS -- shrugs mockery, busy cycles, sleeps well; consistent. +lutek_multi_burstON_007: PASS -- five days easygoing, restless, warm thanks; sane. +lutek_multi_burstON_008: PASS -- warm, restless, mockery ignored; believable day flow. +lutek_multi_burstON_009: PASS -- many kindnesses warmly taken, restless, sleeps; sane. +lutek_multi_burstON_010: PASS -- restless poet, warm, recovers nightly; nothing jars. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..b504575 --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstON_001: PASS -- calm baseline, single justified snap at mockery, sleeps and recovers. +welf_multi_burstON_002: PASS -- even-keeled, lets mockery pass, sane rest/boredom cadence. +welf_multi_burstON_003: PASS -- composed throughout, mockery shrugged off, believable pacing. +welf_multi_burstON_004: PASS -- steady, warm to kindness, mockery ignored, recovers nightly. +welf_multi_burstON_005: PASS -- calm rain-watch tension fades, sane daily rhythm. +welf_multi_burstON_006: PASS -- composed, lets barbs pass, believable boredom-rest cycle. +welf_multi_burstON_007: PASS -- mild rain-driven edge, recovers overnight, drifting wake times. +welf_multi_burstON_008: PASS -- six days steady, mockery shrugged, plausible sleep drift. +welf_multi_burstON_009: PASS -- transactional calm, ignores barbs, hunger/temper rise and fall believably. +welf_multi_burstON_010: PASS -- even-keeled, lets mockery pass, sane recovery and pacing. \ No newline at end of file diff --git a/eval/hourly_runs/slice_0/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_0/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..8b2ab1f --- /dev/null +++ b/eval/hourly_runs/slice_0/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,13 @@ +I have read the full file. I'll judge each record now. + +wojslaw_multi_burstON_001: PASS -- proud, prickly, sleeps/wakes, sane swings. +wojslaw_multi_burstON_002: PASS -- calm warm start, escalating temper, recovers nightly. +wojslaw_multi_burstON_003: PASS -- believable arc, gratitude then friction, sleeps well. +wojslaw_multi_burstON_004: PASS -- prickly flares, calm spells, normal pacing. +wojslaw_multi_burstON_005: PASS -- short temper, on-edge stretches, recovers overnight. +wojslaw_multi_burstON_006: PASS -- sane mix of thanks and eruptions, nightly reset. +wojslaw_multi_burstON_007: PASS -- warm and prickly by turns, believable spans. +wojslaw_multi_burstON_008: PASS -- proud refusals and flares, sleeps and recovers. +wojslaw_multi_burstON_009: PASS -- escalating irritation, calm mornings, sane pacing. + +Note: only 9 records present in the file (001-009); no record 010 exists. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..2fe1c1d --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_011: PASS -- calm productive day, sane pacing, sleeps and wakes recovered +branic_day_burstOFF_012: PASS -- shrugs off mocks, works, rests, sleeps; believable green recruit +branic_day_burstOFF_013: PASS -- snaps then refuses an order after barbs; thin-skinned recruit fits +branic_day_burstOFF_014: PASS -- restless cycles, kindnesses warmly received, sleeps just after midnight +branic_day_burstOFF_015: PASS -- obedient busy day, mild edge late, sleeps and recovers +branic_day_burstOFF_016: PASS -- steady day, warmth and orders handled, sleeps and wakes settled +branic_day_burstOFF_017: PASS -- many kindnesses, brief edge, restless cycles, normal sleep +branic_day_burstOFF_018: PASS -- eager compliance, warm responses, mild evening edge, normal sleep +branic_day_burstOFF_019: PASS -- shrugs repeated mocks, restless cycles, sleeps and wakes settled +branic_day_burstOFF_020: PASS -- calm restless day, warm responses, sleeps just after midnight \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..2cdee30 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_011: PASS -- Sane pacing, sleeps, wakes recovered; mock-fury fits resentment. +cichy_day_burstOFF_012: PASS -- Believable arc; flares at guard mockery, calms, sleeps well. +cichy_day_burstOFF_013: PASS -- Calm day, ignores soup, one snap at mockery; coherent. +cichy_day_burstOFF_014: PASS -- Two furies at mockery, warm to strangers; consistent and bounded. +cichy_day_burstOFF_015: PASS -- Steady day, single fury at barb, recovers and sleeps. +cichy_day_burstOFF_016: PASS -- Two evening furies at mockery; sane rise and fall. +cichy_day_burstOFF_017: PASS -- Warm to strangers, curt/furious to guards; believable pacing. +cichy_day_burstOFF_018: PASS -- Coherent day, furies at mockery, recovers by morning. +cichy_day_burstOFF_019: PASS -- Calm with one snap; bounded temper, normal sleep cycle. +cichy_day_burstOFF_020: PASS -- Quiet day, only curt to a barb; sane throughout. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..672ce9b --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_011: PASS -- calm, composed; warm to kindness, ignores mockery; sleeps well. +edda_day_burstOFF_012: PASS -- composed day; mild complaint at a stranger's order fits authority. +edda_day_burstOFF_013: PASS -- settled and warm; lets a stranger's order pass calmly. +edda_day_burstOFF_014: PASS -- calm throughout; warmth and unruffled responses, normal sleep. +edda_day_burstOFF_015: PASS -- composed, gracious to help, ignores barbs; sane pacing. +edda_day_burstOFF_016: PASS -- mild complaint at stranger's order; otherwise composed and warm. +edda_day_burstOFF_017: PASS -- settled, warm to kindness, lets order pass; normal night. +edda_day_burstOFF_018: PASS -- curt reply and declining a stranger's order fit her authority. +edda_day_burstOFF_019: PASS -- warm, composed, ignores mockery; believable pacing and sleep. +edda_day_burstOFF_020: PASS -- refuses stranger's order, stays composed; fits commanding standing. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..82c2c1c --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_011: PASS -- calm day, warm to kindness, lets mockery pass; sane. +halgrim_day_burstOFF_012: PASS -- cold curtness escalating with Wojsław, refuses arrogance, obeys Edda; consistent. +halgrim_day_burstOFF_013: PASS -- cold to provocation, but warm to Marta; even keel holds. +halgrim_day_burstOFF_014: PASS -- settled throughout, declines Wojsław, obeys Edda; believable. +halgrim_day_burstOFF_015: FLAG -- mutters complaint at Edda, whom he respects; slightly off. +halgrim_day_burstOFF_016: PASS -- uneventful calm day, warm to kindness; entirely sane. +halgrim_day_burstOFF_017: PASS -- steady, obeys Edda, warm to soup; nothing jars. +halgrim_day_burstOFF_018: PASS -- mostly calm, mild curtness late after repeated mockery; fine. +halgrim_day_burstOFF_019: PASS -- settled all day, obeys Edda, warm to kindness; believable. +halgrim_day_burstOFF_020: PASS -- quiet calm day, lets mockery pass, warm to Marta; sane. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..a6346f7 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_011: PASS -- restless cycles, shrugs off mockery, sleeps and wakes calm. +lutek_day_burstOFF_012: PASS -- novelty-seeking rhythm, warm thanks, normal sleep cycle. +lutek_day_burstOFF_013: PASS -- believable restless day, kindnesses welcomed, recovers overnight. +lutek_day_burstOFF_014: PASS -- sane pacing, good-humoured, sleeps and wakes settled. +lutek_day_burstOFF_015: PASS -- mockery shrugged off, restless cycles, late but sane sleep. +lutek_day_burstOFF_016: PASS -- public mockery rolls off, normal restless-rest rhythm. +lutek_day_burstOFF_017: PASS -- calm response to mockery, believable restless day. +lutek_day_burstOFF_018: PASS -- frequent novelty-seeking, warm to kindness, normal night. +lutek_day_burstOFF_019: PASS -- sane restless cycles, kindnesses welcomed, sleeps fine. +lutek_day_burstOFF_020: PASS -- believable pacing, easy-going, sleeps and wakes settled. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..5d0f672 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_011: PASS -- Restlessness cycles, takes mockery in stride, sleeps and wakes recovered. +welf_day_burstOFF_012: PASS -- Sane boredom-rest cycle, warm to kindness, recovers overnight. +welf_day_burstOFF_013: PASS -- Believable pacing, calm, kindness welcomed, sleeps and wakes settled. +welf_day_burstOFF_014: PASS -- Even-keeled day, boredom rises and falls, recovers overnight. +welf_day_burstOFF_015: PASS -- Steady cycle of restlessness and rest, calm throughout. +welf_day_burstOFF_016: PASS -- Normal boredom pacing, settled, no jarring reactions. +welf_day_burstOFF_017: PASS -- Believable day, brief tension passes, calm and recovered. +welf_day_burstOFF_018: PASS -- Sane pacing, mild tension fades, sleeps and recovers. +welf_day_burstOFF_019: PASS -- Calm even day, kindness welcomed, overnight recovery intact. +welf_day_burstOFF_020: PASS -- Steady, warm to kindness, believable boredom cycle, recovers overnight. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..08c999d --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_011: PASS -- proud man flares at mocks, calm otherwise, sleeps and recovers +wojslaw_day_burstOFF_012: PASS -- sane pacing, fury at insults, warmth at kindness, night reset +wojslaw_day_burstOFF_013: PASS -- mostly calm, snaps at public mockery, believable arc +wojslaw_day_burstOFF_014: PASS -- temper builds across afternoon, recovers by night, consistent +wojslaw_day_burstOFF_015: PASS -- repeated mocks provoke fury, kindness warms, sane day +wojslaw_day_burstOFF_016: PASS -- defiant to orders, erupts at insult, recovers normally +wojslaw_day_burstOFF_017: PASS -- calm day, single eruption at public mockery, believable +wojslaw_day_burstOFF_018: PASS -- flares at mocks, lingering tension, sleeps and recovers +wojslaw_day_burstOFF_019: PASS -- calm day, one fury at late mock, sane pacing +wojslaw_day_burstOFF_020: PASS -- defiant and fiery under heavy provocation, plausible for proud man \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_1/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..66a1432 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_011: PASS -- restless/settled cycles sane; one snap at Halgrim fits thin-skinned recruit. +branic_day_burstON_012: PASS -- calm day, lets barbs pass, obeys orders; believable young soldier. +branic_day_burstON_013: PASS -- one fury at evening mocking fits reactive temper; otherwise composed. +branic_day_burstON_014: PASS -- restless-bored cycles, kindness warmly received, late sleep plausible. +branic_day_burstON_015: PASS -- busy obedient day, tension builds, calm sleep; coherent pacing. +branic_day_burstON_016: PASS -- ordinary day, lets mocking pass, warms to kindness; sane. +branic_day_burstON_017: PASS -- many kindnesses warmly taken, orders obeyed, tension ebbs; believable. +branic_day_burstON_018: PASS -- restless cycles, warmth to help, evening edge then calm sleep. +branic_day_burstON_019: PASS -- repeated mocking let pass, obeys, warms to aid; coherent. +branic_day_burstON_020: PASS -- bored-restless rhythm, kindness welcomed, late sleep; sane day. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_1/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..1e99a76 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_011: PASS -- Calm day, curt to guards, warm to stranger; sane. +cichy_day_burstON_012: PASS -- Snaps at mock, otherwise curt; believable edgy captive. +cichy_day_burstON_013: PASS -- Early flare, settles; resentment toward guards consistent. +cichy_day_burstON_014: PASS -- Erupts at public mockery, recovers; fits volatile captive. +cichy_day_burstON_015: PASS -- Curt to guards, calm pacing; sane day. +cichy_day_burstON_016: PASS -- Escalating evening snaps after repeated jabs; believable. +cichy_day_burstON_017: PASS -- Warm to strangers, curt to guards; coherent day. +cichy_day_burstON_018: PASS -- Curt to guards throughout, sleeps; nothing jars. +cichy_day_burstON_019: PASS -- Mild curtness to guards, calm pacing; sane. +cichy_day_burstON_020: PASS -- Mostly settled, curt once to guard; believable. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_1/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..8244e09 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_011: PASS -- calm castellan, warmth to kindness, lets mockery pass +edda_day_burstON_012: PASS -- composed, mild complaint at being ordered, sleeps and recovers +edda_day_burstON_013: PASS -- settled, warm to help, unbothered by stranger's order +edda_day_burstON_014: PASS -- calm day, kindness welcomed, barb ignored, recovers overnight +edda_day_burstON_015: PASS -- composed throughout, warm thanks, mockery passes without reaction +edda_day_burstON_016: PASS -- settled, mild complaint at order, otherwise calm and warm +edda_day_burstON_017: PASS -- calm, gracious to kindness, order let pass, recovers +edda_day_burstON_018: PASS -- curt answer and refusal to stranger's order, in-command +edda_day_burstON_019: PASS -- consistently settled, warm to help, mockery ignored +edda_day_burstON_020: PASS -- authoritative refusal of stranger's order, otherwise calm and warm \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_1/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..0c25b31 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_011: PASS -- calm, warm to kindness, lets Wojsław's jab pass; believable. +halgrim_day_burstON_012: PASS -- cold curtness to Wojsław's mockery and refusals; fitting restraint. +halgrim_day_burstON_013: FLAG -- cold contempt twice at respected Edda's orders jars. +halgrim_day_burstON_014: PASS -- even-keeled, declines Wojsław, warm to kindness; sane day. +halgrim_day_burstON_015: PASS -- mild curtness and a complaint late; recovers overnight. +halgrim_day_burstON_016: PASS -- placid, warm to soup and help throughout; believable. +halgrim_day_burstON_017: PASS -- compliant with Edda, settled all day; sane. +halgrim_day_burstON_018: PASS -- absorbs jabs, one cold reply, mild edge by night. +halgrim_day_burstON_019: PASS -- calm, obliges Edda, warmed by kindness; consistent. +halgrim_day_burstON_020: PASS -- placid throughout, lets Wojsław's barb pass; fitting. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_1/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..f3cbc06 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_011: PASS -- restless cycles, mockery shrugged off, sleeps and wakes at ease +lutek_day_burstON_012: PASS -- believable boredom-rest rhythm, warm to kindness, calm throughout +lutek_day_burstON_013: PASS -- sane pacing, kindnesses welcomed, no jarring reactions +lutek_day_burstON_014: PASS -- normal day, well-fed, restless cycles, sleeps and recovers +lutek_day_burstON_015: PASS -- mockery shrugged off, restless rhythm, late but believable sleep +lutek_day_burstON_016: PASS -- mockery in front of others ignored, calm restless day +lutek_day_burstON_017: PASS -- believable boredom and rest cycles, mockery rolls off +lutek_day_burstON_018: PASS -- warm to many kindnesses, settled, sane daily rhythm +lutek_day_burstON_019: PASS -- restless cycles, fed, mild edge then calm, sleeps fine +lutek_day_burstON_020: PASS -- normal fed day, restless rhythm, sleeps and wakes restless \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_1/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..c6d17f7 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_011: PASS -- restless cycles, warm to kindness, flares only at mockery +welf_day_burstON_012: PASS -- believable boredom-rest cycle, calm, warm to help +welf_day_burstON_013: PASS -- sane pacing, settled and grateful throughout +welf_day_burstON_014: PASS -- even-keeled day, restless then occupied, sleeps recovered +welf_day_burstON_015: PASS -- calm composed day, boredom rises and eases +welf_day_burstON_016: PASS -- restless cycles, brief edge, no unprovoked anger +welf_day_burstON_017: PASS -- busy restless day, warm, momentarily tense, recovers +welf_day_burstON_018: PASS -- believable boredom rhythm, calm, sleeps and wakes settled +welf_day_burstON_019: PASS -- steady composed day, grateful, sane pacing +welf_day_burstON_020: PASS -- many kindnesses well-received, calm, no jarring beats \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_1/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..eda8002 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_011: PASS -- Flares at mocks, warm to kindness, sleeps and wakes recovered. +wojslaw_day_burstON_012: PASS -- Believable prickly day, refusals and eruptions, calm night reset. +wojslaw_day_burstON_013: PASS -- Mostly calm day with proportionate flares; sane pacing. +wojslaw_day_burstON_014: PASS -- Temper builds over day, eruptions fit; recovers overnight. +wojslaw_day_burstON_015: PASS -- Hot day, many eruptions but plausible for prickly noble. +wojslaw_day_burstON_016: PASS -- Mix of warmth, refusals, eruptions; coherent pacing and reset. +wojslaw_day_burstON_017: PASS -- Calm day, occasional flares; believable and recovers overnight. +wojslaw_day_burstON_018: PASS -- Sustained edginess after eruptions, plausible; sleeps and recovers. +wojslaw_day_burstON_019: PASS -- Mostly calm, single late eruption; sane day and reset. +wojslaw_day_burstON_020: PASS -- Very hot day, many refusals/eruptions; plausible for prickly noble. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..e612166 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows sane day/night cadence, restlessness when idle, warmth at kindness, escalating temper toward Halgrim's barbs in the evening, recovery after sleep — consistent with the raw, easily-bored, thin-skinned recruit profile. No absurd pacing or contradictions. + +branic_multi_burstOFF_011: PASS -- sane cadence; idle-restlessness, evening flares, sleep recovery fit profile. +branic_multi_burstOFF_012: PASS -- believable pacing; mild flares, defiance, nightly reset all fit. +branic_multi_burstOFF_013: PASS -- compliant-then-fraying arc, kindness warmth, sleep recovery all sane. +branic_multi_burstOFF_014: PASS -- calm obedient stretch with late flares; pacing fully believable. +branic_multi_burstOFF_015: PASS -- coherent days; rain/mock-driven defiance and recovery fit profile. +branic_multi_burstOFF_016: PASS -- sane cadence; escalating refusals and evening eruptions fit recruit. +branic_multi_burstOFF_017: PASS -- believable; morning flares after edge, warmth at kindness, recovers. +branic_multi_burstOFF_018: PASS -- coherent; clustered Halgrim friction and nightly reset all sane. +branic_multi_burstOFF_019: PASS -- calm obliging days with occasional flare; pacing fully believable. +branic_multi_burstOFF_020: PASS -- sane days; rain-day defiance and eruptions fit thin-skinned recruit. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..da0903f --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,12 @@ +All records show consistent sane patterns: night sleep with recovery, escalation to provocations (fury at mocks), curt/neutral toward soup, warmth to strangers, tense/settled cycling. Profile fit is good. Nothing absurd. + +cichy_multi_burstOFF_011: PASS -- Sane cycles; resents guards, fury at mocks, recovers nightly. +cichy_multi_burstOFF_012: PASS -- Believable pacing; provocation-driven fury, warm to strangers. +cichy_multi_burstOFF_013: PASS -- Calm baseline, sane escalation to mocks, nightly reset. +cichy_multi_burstOFF_014: PASS -- Coherent days; tense under rain, fury at provocations. +cichy_multi_burstOFF_015: PASS -- Many provocations but proportionate fury, recovers each night. +cichy_multi_burstOFF_016: PASS -- Warm to strangers, fury at guards, sane rhythm. +cichy_multi_burstOFF_017: PASS -- Heavy mocking clustered believably; settles between, sleeps nightly. +cichy_multi_burstOFF_018: PASS -- Calm baseline, fury only at mocks, nightly recovery. +cichy_multi_burstOFF_019: PASS -- Rain tension, provocation fury, warmth to strangers, sane reset. +cichy_multi_burstOFF_020: PASS -- Believable arc; resents guards, fury at mocks, recovers. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..a11921d --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,12 @@ +All records read. Each shows composed, in-command behavior: warm to kindness, firm/curt to provocations, sleeps at night, wakes settled. Reactions scale sensibly and stay within profile. + +edda_multi_burstOFF_011: PASS -- composed, warm to kindness, sleeps nightly, nothing jars +edda_multi_burstOFF_012: PASS -- curt to mocks, sane recovery, in-command throughout +edda_multi_burstOFF_013: PASS -- calm, warm responses, firm refusals, believable pacing +edda_multi_burstOFF_014: PASS -- composed, lets provocations pass, sleeps and recovers +edda_multi_burstOFF_015: PASS -- steady, authoritative, sane daily rhythm +edda_multi_burstOFF_016: PASS -- warm and settled, firm with orders, nothing absurd +edda_multi_burstOFF_017: PASS -- composed, refuses orders firmly, recovers nightly +edda_multi_burstOFF_018: PASS -- calm, lets jabs pass, in-command throughout +edda_multi_burstOFF_019: PASS -- curt to mocks, defiant refusals, sane pacing +edda_multi_burstOFF_020: PASS -- composed, warm, firm with orders, recovers each night \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..2142364 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_011: PASS -- calm baseline, cold curtness escalates under Wojsław's mocking, recovers nightly. +halgrim_multi_burstOFF_012: PASS -- steady, mild curtness to provocation, sleeps and recovers believably. +halgrim_multi_burstOFF_013: PASS -- even keel, tension builds and resets across nights sensibly. +halgrim_multi_burstOFF_014: PASS -- cold defiance to Wojsław, warm to kindness, recovers nightly. +halgrim_multi_burstOFF_015: PASS -- coherent arc, tension from provocation, warmth otherwise, nightly reset. +halgrim_multi_burstOFF_016: PASS -- calm throughout, late cold contempt to Wojsław, sane pacing. +halgrim_multi_burstOFF_017: PASS -- rain plus mockery wears him to curtness, recovers nightly. +halgrim_multi_burstOFF_018: PASS -- steady, cold responses to Wojsław, tension fades with sleep. +halgrim_multi_burstOFF_019: PASS -- calm, compliant, mild curtness, believable spans and recovery. +halgrim_multi_burstOFF_020: PASS -- even-keeled, declines Wojsław coolly, warm otherwise, nightly reset. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..bc8deb7 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,12 @@ +All records show consistent sane behavior: restless/occupies cycle, mockery rolls off (occasional mild curt reply, fits thick-skinned), warm thanks for kindness, sleeps at night, wakes settled. Nothing jars. + +lutek_multi_burstOFF_011: PASS -- sane restless-rest cycle, mockery shrugged, sleeps and recovers +lutek_multi_burstOFF_012: PASS -- believable pacing, warm to kindness, restful nights +lutek_multi_burstOFF_013: PASS -- mockery passes, restless cycle, good sleep recovery +lutek_multi_burstOFF_014: PASS -- one mild curt reply fits thick-skinned, otherwise calm +lutek_multi_burstOFF_015: PASS -- consistent restless-rest rhythm, warm, sleeps well +lutek_multi_burstOFF_016: PASS -- sane days, mockery shrugged, recovers overnight +lutek_multi_burstOFF_017: PASS -- repeated mockery rolls off, restless cycle, good sleep +lutek_multi_burstOFF_018: PASS -- believable pacing, mockery passes, restful nights +lutek_multi_burstOFF_019: PASS -- warm, restless-rest cycle, sleeps and wakes settled +lutek_multi_burstOFF_020: PASS -- sane three days, easy-going, recovers overnight \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..ec77d98 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,12 @@ +All records show consistent sane pacing: boredom rises and is relieved, rests, sleeps at night, wakes recovered, takes kindness warmly, shrugs off mockery (consistent with calm Welf). One brief bristle in 016 (single mock) is minor and recovers. No absurdities. + +welf_multi_burstOFF_011: PASS -- sane idle-boredom cycles, sleeps and recovers, shrugs off mockery +welf_multi_burstOFF_012: PASS -- believable boredom/rest cycles, calm under mockery, nightly recovery +welf_multi_burstOFF_013: PASS -- coherent days, settles and rests, calm responses throughout +welf_multi_burstOFF_014: PASS -- steady boredom-relief rhythm, sleeps well, no jarring reactions +welf_multi_burstOFF_015: PASS -- sane pacing, calm under barbs, recovers each night +welf_multi_burstOFF_016: PASS -- one brief bristle at mockery, recovers; otherwise calm and coherent +welf_multi_burstOFF_017: PASS -- believable idle cycles, shrugs off barbs, nightly reset +welf_multi_burstOFF_018: PASS -- consistent boredom/rest pattern, even-keeled throughout +welf_multi_burstOFF_019: PASS -- sane multi-day rhythm, calm responses, sleeps and recovers +welf_multi_burstOFF_020: PASS -- coherent pacing, warm to kindness, unruffled by mockery \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..7069109 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows believable pacing, night sleep with recovery, and a hot-tempered proud man who erupts at mocks but warms to kindness. Nothing absurd. + +wojslaw_multi_burstOFF_011: PASS -- calm baseline, erupts at public mocks, recovers nightly +wojslaw_multi_burstOFF_012: PASS -- believable flares at insults, settles between, sleeps well +wojslaw_multi_burstOFF_013: PASS -- escalating mock-days, eruptions proportionate, recovers each morning +wojslaw_multi_burstOFF_014: PASS -- heavy provocation, hot defiant reactions fit proud temper +wojslaw_multi_burstOFF_015: PASS -- sane arc, anger tracks mocks, nightly reset holds +wojslaw_multi_burstOFF_016: PASS -- many insults, fitting fury and refusals, recovers overnight +wojslaw_multi_burstOFF_017: PASS -- believable temper across six days, sleeps and recovers +wojslaw_multi_burstOFF_018: PASS -- provocation-heavy but proportionate; warms to kindness when calm +wojslaw_multi_burstOFF_019: PASS -- warm early, sharpens with mocks, plausible throughout +wojslaw_multi_burstOFF_020: PASS -- prickly eruptions at insults, calms between, nightly reset \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..65811e6 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstON_branic.txt @@ -0,0 +1,23 @@ +All records read. Judging each: sane day pacing, sleeps at night, recovers, restless/bored, warms to kindness, snaps at Halgrim's mocking when worn down — all consistent with a reactive young recruit. Checking each for absurdities. + +branic_multi_burstON_011: PASS -- sane pacing, defiance builds, sleep resets, fits profile +branic_multi_burstON_012: PASS -- believable, mostly compliant with occasional eruption +branic_multi_burstON_013: snaps at Marta's soup (D5 07:56) — mild oddity but isolated, on a worn rainy morning. Borderline but acceptable as displaced. +branic_multi_burstON_014: PASS -- calm stretches, late eruptions, recovers nightly +branic_multi_burstON_015: PASS -- restless, refusals cluster under strain, sleeps fine +branic_multi_burstON_016: PASS -- believable boredom/temper rhythm across days +branic_multi_burstON_017: PASS -- mostly compliant, brief morning flare, sane +branic_multi_burstON_018: PASS -- escalating Halgrim friction Day6, nightly reset, fits +branic_multi_burstON_019: PASS -- calm, warm to kindness, isolated late eruptions +branic_multi_burstON_020: PASS -- sane pacing, refusals under strain, recovers + +branic_multi_burstON_011: PASS -- sane pacing, sleeps/recovers, temper fits reactive recruit +branic_multi_burstON_012: PASS -- believable days, mostly compliant with rare eruption +branic_multi_burstON_013: FLAG -- snaps at Marta's kind soup (Day5 07:56) +branic_multi_burstON_014: PASS -- calm stretches, late flares, nightly reset +branic_multi_burstON_015: PASS -- restless, refusals cluster under strain, sane +branic_multi_burstON_016: PASS -- believable boredom and temper rhythm +branic_multi_burstON_017: PASS -- mostly compliant, brief morning flare, sane +branic_multi_burstON_018: PASS -- escalating Halgrim friction, recovers each night +branic_multi_burstON_019: PASS -- calm, warm to kindness, isolated eruptions +branic_multi_burstON_020: PASS -- sane pacing, refusals under strain, recovers \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..33335d9 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,12 @@ +All records read. Judging each on its own for sane pacing and consistency with a resentful, on-edge captive who escalates at guards' mocking but stays calm-to-curt otherwise, sleeps at night and wakes recovered. + +cichy_multi_burstON_011: PASS -- sane arc; mockery escalates, soup tolerated, nightly recovery +cichy_multi_burstON_012: PASS -- curt-to-cold at mockery, calm baseline, recovers nightly +cichy_multi_burstON_013: PASS -- believable tension cycles, soup curt, sleeps and recovers +cichy_multi_burstON_014: PASS -- escalation on mockery, rain stressor, recovers each night +cichy_multi_burstON_015: PASS -- repeated mockery builds to fury believably, recovers overnight +cichy_multi_burstON_016: PASS -- warms to strangers, fury at public mockery, nightly reset +cichy_multi_burstON_017: PASS -- heavy mockery day builds fury, otherwise calm, recovers +cichy_multi_burstON_018: PASS -- soup once bristles when tense; plausible, recovers nightly +cichy_multi_burstON_019: PASS -- curt baseline, rain stressor, sane cycles, nightly recovery +cichy_multi_burstON_020: PASS -- calm baseline, curt at mockery/soup, sleeps and recovers \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..4101918 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_011: PASS -- calm, composed, warm to kindness; sane sleep cycle +edda_multi_burstON_012: PASS -- mild escalation to provocation by day 6, recovers nightly +edda_multi_burstON_013: PASS -- composed authority, declines orders, warm otherwise +edda_multi_burstON_014: PASS -- settled days, sensible meals and rest, sleeps well +edda_multi_burstON_015: PASS -- firm with strangers' orders, warm to staff, sane pacing +edda_multi_burstON_016: PASS -- consistently composed, warm, good day/night rhythm +edda_multi_burstON_017: PASS -- refuses strangers' orders firmly, otherwise settled and warm +edda_multi_burstON_018: PASS -- mostly settled, brief edge day 4, recovers overnight +edda_multi_burstON_019: PASS -- curt to mockery, refuses orders, resets each night +edda_multi_burstON_020: PASS -- calm composed castellan, sensible rhythm throughout \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..db86fa1 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_011: PASS -- calm, cold curtness under Wojsław's needling; sleeps and recovers. +halgrim_multi_burstON_012: PASS -- steady, controlled curtness escalating only with provocation; nightly reset. +halgrim_multi_burstON_013: PASS -- even-keeled, tension builds and settles believably over days. +halgrim_multi_burstON_014: PASS -- cold defiance toward Wojsław, warmth to Marta; sane pacing. +halgrim_multi_burstON_015: PASS -- controlled curtness, recovers nightly; provocations build tension plausibly. +halgrim_multi_burstON_016: PASS -- mostly settled, cold only when pushed; believable arc. +halgrim_multi_burstON_017: PASS -- rain and needling wear him; cold curtness, nightly recovery. +halgrim_multi_burstON_018: PASS -- steady, cold contempt under mockery; sane day rhythm. +halgrim_multi_burstON_019: PASS -- calm throughout, mild curtness at most; well-regulated. +halgrim_multi_burstON_020: PASS -- composed across six days, warmth and measured curtness; recovers nightly. \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..aa0f395 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,12 @@ +All records show consistent sane patterns: restless/seeking cycles, warm reactions to kindness, mockery rolled off (occasional mild curt reply fits a thick-skinned man having one off moment), nightly sleep with recovery. Nothing jars. + +lutek_multi_burstON_011: PASS -- restless-seeking poet, warm to kindness, mockery shrugged off +lutek_multi_burstON_012: PASS -- sane cycles, easy gratitude, sleeps and recovers nightly +lutek_multi_burstON_013: PASS -- mockery ignored, restless tinkering, believable daily pacing +lutek_multi_burstON_014: PASS -- one curt reply fits; otherwise calm, restless, recovering +lutek_multi_burstON_015: PASS -- warm, seeking novelty, sane sleep-wake rhythm throughout +lutek_multi_burstON_016: PASS -- mockery shrugged, restless cycles, believable nights and recovery +lutek_multi_burstON_017: PASS -- one cold retort isolated; otherwise warm, restless, sane +lutek_multi_burstON_018: PASS -- mockery let pass, restless tinkering, normal day rhythm +lutek_multi_burstON_019: PASS -- warm to kindness, seeks stimulus, sleeps and recovers +lutek_multi_burstON_020: PASS -- calm restless poet, warm thanks, believable pacing throughout \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..19cd080 --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstON_welf.txt @@ -0,0 +1,12 @@ +All records reviewed. Most show calm Welf cycling boredom/occupation/rest, settled mood, sleeping at night, letting mocks pass. Record 016 has an outburst at Day 3 06:58 ("flares up and snaps angrily") plus bristling — out of character for a composed, even-keeled man whose trouble is idleness not people. + +welf_multi_burstON_011: PASS -- calm restless-occupy cycle, sleeps nightly, lets mock pass +welf_multi_burstON_012: PASS -- sane boredom/rest cycle, brief rain-edge tension, mocks ignored +welf_multi_burstON_013: PASS -- believable pacing, recovers nightly, neutral to barbs +welf_multi_burstON_014: PASS -- steady idleness cycle, calm, sleeps and wakes recovered +welf_multi_burstON_015: PASS -- composed throughout, rain edge mild, lets mock pass +welf_multi_burstON_016: FLAG -- Day3 snaps angrily/bristles at mocks; jars for even-keeled Welf +welf_multi_burstON_017: PASS -- calm restless cycle, mild morning tension, mocks shrugged off +welf_multi_burstON_018: PASS -- believable days, recovers nightly, neutral to barb +welf_multi_burstON_019: PASS -- sane pacing across six days, calm, sleeps nightly +welf_multi_burstON_020: PASS -- composed throughout, rain tension mild, lets mocks pass \ No newline at end of file diff --git a/eval/hourly_runs/slice_1/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_1/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..7e6e14f --- /dev/null +++ b/eval/hourly_runs/slice_1/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows sane day/night cycles, recovery on sleep, anger driven by mocks/orders consistent with a proud hot-tempered noble, kindness/soup warmly received. Checking for anything jarring: erupting at soup/kindness occurs when already enraged (carry-over irritability), believable. No starvation, no causeless rage. + +wojslaw_multi_burstON_011: PASS -- calm baseline, occasional eruptions on insults; recovers nightly +wojslaw_multi_burstON_012: PASS -- prickly, flares at mocks/orders; sane pacing and sleep +wojslaw_multi_burstON_013: PASS -- escalates over days but resets each morning; believable +wojslaw_multi_burstON_014: PASS -- hot day of insults, heavy refusals; fits proud temper +wojslaw_multi_burstON_015: PASS -- mounting irritation, soup-anger from carry-over; nightly recovery +wojslaw_multi_burstON_016: PASS -- dense eruptions but plausibly insult-driven; sleeps and resets +wojslaw_multi_burstON_017: PASS -- mixed warmth and fury; coherent rise/fall across days +wojslaw_multi_burstON_018: PASS -- volatile but tied to mocks; recovers each night +wojslaw_multi_burstON_019: PASS -- warm calm days, late-day flares on insults; believable +wojslaw_multi_burstON_020: PASS -- prickly throughout, anger keyed to provocations; sane sleep \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..2d33726 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_021: PASS -- sane pacing, boredom cycles, sleeps and wakes recovered +branic_day_burstOFF_022: PASS -- believable day, warm to kindness, restful night +branic_day_burstOFF_023: PASS -- coherent boredom/rest rhythm, recovers overnight +branic_day_burstOFF_024: PASS -- consistent warmth and restlessness, normal sleep cycle +branic_day_burstOFF_025: PASS -- sane day, mild tension, recovers after sleep +branic_day_burstOFF_026: PASS -- believable rhythm, warm responses, settled overnight +branic_day_burstOFF_027: PASS -- coherent pacing, obeys orders, sleeps and recovers +branic_day_burstOFF_028: PASS -- normal duty cycle, compliant, restful night +branic_day_burstOFF_029: PASS -- sane day, warm to soup, recovers overnight +branic_day_burstOFF_030: PASS -- believable rhythm, obeys orders, settled after sleep \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..a39ddcc --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_021: PASS -- sensible day; mocks anger, soup curtness, sleeps and wakes calm +cichy_day_burstOFF_022: PASS -- believable pacing; snaps at mockery, warm to stranger, recovers +cichy_day_burstOFF_023: PASS -- calm stretches with curt soup and one furious eruption +cichy_day_burstOFF_024: PASS -- coherent day; fury at mocks, curt soup, rests and sleeps +cichy_day_burstOFF_025: PASS -- mocks provoke fury, soup bristle, settles overnight, fine +cichy_day_burstOFF_026: PASS -- quiet day; no provocation, no anger, sane pacing +cichy_day_burstOFF_027: PASS -- mostly calm, one snap at mockery, recovers overnight +cichy_day_burstOFF_028: PASS -- graded reactions to mocks, calm intervals, sleeps recovered +cichy_day_burstOFF_029: PASS -- repeated mocks bring fury, soup curtness, calm between +cichy_day_burstOFF_030: PASS -- warm to stranger, curt soup, fury at mock, sleeps \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..1b9e61a --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_021: PASS -- calm warmth, mild complaint at stranger's order; sane day. +edda_day_burstOFF_022: PASS -- composed, warm to kindness, lets stranger's order pass. +edda_day_burstOFF_023: PASS -- unbothered by mockery and orders, settled throughout. +edda_day_burstOFF_024: PASS -- refuses stranger's order, warm otherwise; in-command and sane. +edda_day_burstOFF_025: PASS -- steady warmth, restful pacing, sleeps and wakes recovered. +edda_day_burstOFF_026: PASS -- shrugs off mockery and orders, settled and composed. +edda_day_burstOFF_027: PASS -- lets mockery pass, refuses an order; consistent and sane. +edda_day_burstOFF_028: PASS -- complies once, warm throughout; believable composed day. +edda_day_burstOFF_029: PASS -- uniformly warm to kindnesses, restful, sleeps normally. +edda_day_burstOFF_030: PASS -- quiet steady day, warm to soup, settled. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..88d12ff --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_021: PASS -- calm, warm to kindness, ignores Wojsław's jab, sleeps well. +halgrim_day_burstOFF_022: PASS -- obeys Edda, declines Wojsław, even-keeled throughout. +halgrim_day_burstOFF_023: PASS -- settled day, lets mockery pass, follows Edda's order. +halgrim_day_burstOFF_024: PASS -- one ignored Edda order minor; otherwise steady and warm. +halgrim_day_burstOFF_025: PASS -- repeated Wojsław provocation yields curtness and edge, never rage. +halgrim_day_burstOFF_026: PASS -- declines/refuses Wojsław, obeys Edda, stays settled. +halgrim_day_burstOFF_027: PASS -- mostly steady; one mutter at Edda slightly off but minor. +halgrim_day_burstOFF_028: PASS -- placid, warm, lets Wojsław's barbs pass twice. +halgrim_day_burstOFF_029: PASS -- calm throughout, ignores mockery, warm to kindnesses. +halgrim_day_burstOFF_030: PASS -- obeys Edda, curt to Wojsław, even and recovered. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..b2d1fe1 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_021: PASS -- restless cycles, warm to kindness, brief edge, sleeps and recovers +lutek_day_burstOFF_022: PASS -- believable boredom cycles, warm thanks, calm, recovers overnight +lutek_day_burstOFF_023: PASS -- restless-rest rhythm, warm to help, sane late bedtime +lutek_day_burstOFF_024: PASS -- mockery shrugged off, restless cycles, sleeps and wakes recovered +lutek_day_burstOFF_025: PASS -- insults roll off, warm, restless, recovers overnight +lutek_day_burstOFF_026: PASS -- calm warm day, boredom cycles, normal sleep recovery +lutek_day_burstOFF_027: PASS -- mockery ignored, restless rhythm, mild evening edge, recovers +lutek_day_burstOFF_028: PASS -- believable cycles, warm thanks, brief edge, overnight recovery +lutek_day_burstOFF_029: PASS -- mockery shrugged off, warm, restless, sleeps and recovers +lutek_day_burstOFF_030: PASS -- insult ignored, calm warm day, normal sleep recovery \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..255d535 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_021: PASS -- calm, occupied, restful nights, warm to kindnesses +welf_day_burstOFF_022: PASS -- restlessness rises and eases, brief edge believable +welf_day_burstOFF_023: PASS -- sane cycle of boredom, occupation, rest +welf_day_burstOFF_024: PASS -- lets mockery pass, fitting his composed nature +welf_day_burstOFF_025: PASS -- believable boredom-rest rhythm, mild evening tension +welf_day_burstOFF_026: PASS -- steady day, warm to help, sleeps well +welf_day_burstOFF_027: PASS -- even-keeled, restlessness and rest in balance +welf_day_burstOFF_028: PASS -- normal pacing, mild edge, recovers overnight +welf_day_burstOFF_029: PASS -- shrugs off mockery, otherwise calm and occupied +welf_day_burstOFF_030: PASS -- composed throughout, warm responses, sound night \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..fe40e25 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_021: PASS -- flares at mockery, calms, sleeps and wakes recovered +wojslaw_day_burstOFF_022: PASS -- proportionate snaps at insults, warmth to kindness, sane pacing +wojslaw_day_burstOFF_023: PASS -- public mockery erupts, settles between, believable day +wojslaw_day_burstOFF_024: PASS -- flares at mockery, tolerates kindness, normal rhythm +wojslaw_day_burstOFF_025: PASS -- escalating mockery erupts, recovers, sane day arc +wojslaw_day_burstOFF_026: PASS -- mostly warm day, bristles at late mockery, fine +wojslaw_day_burstOFF_027: PASS -- proud refusals of orders, flares at mockery, sane +wojslaw_day_burstOFF_028: PASS -- defiant to orders, erupts at insults, recovers nightly +wojslaw_day_burstOFF_029: PASS -- refuses orders, flares at Marta's mockery, settles +wojslaw_day_burstOFF_030: PASS -- public mockery erupts early, calms across day, sleeps \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_2/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..05decf4 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_021: PASS -- sane day, restless and bored, lets barbs pass +branic_day_burstON_022: PASS -- calm productive day, warm to kindnesses, sleeps well +branic_day_burstON_023: PASS -- believable bored-recruit rhythm, recovers overnight +branic_day_burstON_024: PASS -- steady day, many kindnesses warmly received +branic_day_burstON_025: FLAG -- erupts in fury at a stranger's kindness, jarring +branic_day_burstON_026: PASS -- ordinary busy day, takes mockery in stride +branic_day_burstON_027: PASS -- quiet bored day, mild tension by evening +branic_day_burstON_028: PASS -- dense orders day, obeys, sleeps and recovers +branic_day_burstON_029: PASS -- normal restless day, lets repeated mockery pass +branic_day_burstON_030: PASS -- bristles at public mockery, plausible thin-skinned reaction \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_2/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..432e364 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_021: PASS -- escalating mocks, calm baseline, sleeps and recovers; coherent. +cichy_day_burstON_022: PASS -- curt and angry to mocks, soft to stranger, sane pacing. +cichy_day_burstON_023: PASS -- mostly curt, calm baseline, sensible rise and fall. +cichy_day_burstON_024: PASS -- public mock triggers fury, otherwise measured; believable day. +cichy_day_burstON_025: PASS -- consistently cold and curt to guards; coherent throughout. +cichy_day_burstON_026: PASS -- calm day, curt to soup-bearing guard; recovers overnight. +cichy_day_burstON_027: PASS -- curt then snaps at evening mock; sane pacing. +cichy_day_burstON_028: PASS -- graded response, public mock erupts; recovers, believable. +cichy_day_burstON_029: PASS -- two early eruptions then settling; resentful, coherent arc. +cichy_day_burstON_030: PASS -- warm to stranger, erupts at guard mock; sane. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_2/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..3e042b3 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_021: PASS -- calm composed castellan, sane pacing, sleeps and wakes recovered +edda_day_burstON_022: PASS -- settled day, warm to kindness, lets a stranger's order pass +edda_day_burstON_023: PASS -- composed, ignores mockery and orders, takes charge at night +edda_day_burstON_024: PASS -- calm day, refuses a stranger's order, fitting her authority +edda_day_burstON_025: PASS -- warm to kindnesses, settled throughout, normal sleep cycle +edda_day_burstON_026: PASS -- composed, shrugs off mockery and order, sane pacing +edda_day_burstON_027: PASS -- lets repeated mockery pass, refuses order, stays composed +edda_day_burstON_028: PASS -- settled day, complies once, warm to staff kindnesses +edda_day_burstON_029: PASS -- warm to many kindnesses, calm, takes charge nightly +edda_day_burstON_030: PASS -- quiet settled day, normal meals, sleeps and wakes \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_2/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..08afadf --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_021: PASS -- calm, warm to kindness, ignores Wojsław's jab, sleeps well. +halgrim_day_burstON_022: PASS -- obeys Edda, refuses Wojsław, even keel throughout. +halgrim_day_burstON_023: PASS -- settled day, lets mockery pass, recovers overnight. +halgrim_day_burstON_024: PASS -- one ignored Edda order minor but plausible, otherwise sane. +halgrim_day_burstON_025: PASS -- cold curt replies to repeated Wojsław jabs, fitting. +halgrim_day_burstON_026: PASS -- refuses Wojsław, obeys Edda, steady throughout. +halgrim_day_burstON_027: PASS -- mutters at Edda once is mild, otherwise consistent. +halgrim_day_burstON_028: PASS -- calm, warm, lets Wojsław's jabs pass, sleeps soundly. +halgrim_day_burstON_029: PASS -- unruffled by mockery, warm to kindness, even keel. +halgrim_day_burstON_030: PASS -- cold curtness to Wojsław, obeys Edda, believable arc. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_2/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..96ddc5b --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstON_lutek.txt @@ -0,0 +1,16 @@ +Judging each record on its own against the profile (curious, restless, hard to offend, thick-skinned, easy to amuse). + +Record 021 shows two snaps: he "flares up and snaps angrily" at a mock (19:35) and answers a kindness "curtly and coldly" (22:48). For a thick-skinned man mockery rolls off — flaring up jars with the profile. + +The rest show restlessness, warm thanks, mockery shrugged off, sleep at night, recovery by morning — all sane and in-character. + +lutek_day_burstON_021: FLAG -- flares up at mockery; cold to kindness, contradicts thick-skinned profile +lutek_day_burstON_022: PASS -- restless, warm, sane pacing, sleeps and recovers +lutek_day_burstON_023: PASS -- curious and easy; warm thanks; believable rhythm +lutek_day_burstON_024: PASS -- shrugs off mockery; restless and good-humoured throughout +lutek_day_burstON_025: PASS -- mockery rolls off; warm, restless, normal day +lutek_day_burstON_026: PASS -- easy, restless, warm; sane day and night +lutek_day_burstON_027: PASS -- shrugs mockery; restless, good-humoured, recovers overnight +lutek_day_burstON_028: PASS -- warm, restless, sane pacing and recovery +lutek_day_burstON_029: PASS -- mockery let pass; warm and restless throughout +lutek_day_burstON_030: PASS -- mockery shrugged off; restless, warm, believable \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_2/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..f758d79 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_021: PASS -- calm, occupied, sleeps and wakes normally +welf_day_burstON_022: PASS -- restlessness and ease cycle believably; brief edge fine +welf_day_burstON_023: PASS -- steady day, mild evening tension, recovers overnight +welf_day_burstON_024: PASS -- lets mockery pass, consistent with composed merchant +welf_day_burstON_025: PASS -- boredom-occupy-rest cycles sane; minor evening tension +welf_day_burstON_026: PASS -- even-keeled, warm to kindness, normal sleep +welf_day_burstON_027: PASS -- believable boredom rhythm, mild tension before sleep +welf_day_burstON_028: PASS -- restless-occupy-rest cycling, calm reactions throughout +welf_day_burstON_029: PASS -- shrugs off mockery early, calm steady day +welf_day_burstON_030: PASS -- composed, warm to help, normal day-night cycle \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_2/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..327e93d --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_021: PASS -- proud man flares at insults, calm spells, sleeps and recovers +wojslaw_day_burstON_022: PASS -- believable temper bursts, warm to kindness, sane day rhythm +wojslaw_day_burstON_023: FLAG -- erupts at soup-bringing repeatedly; escalating anger to neutral acts +wojslaw_day_burstON_024: PASS -- sensible flares at mockery, calm recovery, normal pacing +wojslaw_day_burstON_025: PASS -- temper rises and falls believably, sleeps recovered +wojslaw_day_burstON_026: PASS -- mostly calm, snaps late at mockery, coherent day +wojslaw_day_burstON_027: PASS -- proud refusals of orders, flares at mockery, recovers overnight +wojslaw_day_burstON_028: PASS -- defiant refusals and eruptions fit proud temper, sane pacing +wojslaw_day_burstON_029: PASS -- refuses orders, flares then settles, believable arc +wojslaw_day_burstON_030: PASS -- erupts at insults, some displaced anger, recovers by night \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..6b53d7a --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,12 @@ +All records read. Judging each on sane pacing and consistency with a reactive, easily bored young recruit. + +branic_multi_burstOFF_021: PASS -- sane pacing, restless cycles, sleeps, single justified eruption +branic_multi_burstOFF_022: PASS -- believable; refusals/eruptions build under sharp words and rain +branic_multi_burstOFF_023: PASS -- bored cycles, escalating defiance to Halgrim, recovers overnight +branic_multi_burstOFF_024: PASS -- coherent days, mounting refusals after mockery, nightly reset +branic_multi_burstOFF_025: PASS -- restless, warms to kindness, flares at repeated taunts believably +branic_multi_burstOFF_026: PASS -- sane; defiance clusters after barbs then settles +branic_multi_burstOFF_027: PASS -- believable boredom and temper arc across six days +branic_multi_burstOFF_028: PASS -- calm, compliant stretch; warm to kindness, sane pacing +branic_multi_burstOFF_029: PASS -- mostly settled, occasional flare, recovers; believable throughout +branic_multi_burstOFF_030: PASS -- sane days, one public-mockery flare, otherwise calm \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..af38cc4 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstOFF_021: PASS -- sane pacing; resents guards, sleeps and recovers nightly. +cichy_multi_burstOFF_022: PASS -- believable arc; warm to strangers, hostile to guards. +cichy_multi_burstOFF_023: PASS -- coherent days, calm baseline punctuated by guard-triggered anger. +cichy_multi_burstOFF_024: PASS -- consistent resentment, recovers each night, paced well. +cichy_multi_burstOFF_025: PASS -- believable; warms to strangers, snaps at guards. +cichy_multi_burstOFF_026: PASS -- sane pacing, guard provocations escalate appropriately. +cichy_multi_burstOFF_027: PASS -- coherent; many guard frictions but recovers nightly. +cichy_multi_burstOFF_028: PASS -- believable rhythm, resentful toward guards throughout. +cichy_multi_burstOFF_029: PASS -- sane days, kindness welcomed, guards resented. +cichy_multi_burstOFF_030: PASS -- coherent week, consistent grievance, nightly recovery. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..7f9f8da --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_021: PASS -- composed, in-command, sleeps and recovers; sane pacing throughout. +edda_multi_burstOFF_022: PASS -- calm authority, mild complaint at orders fits; believable arc. +edda_multi_burstOFF_023: PASS -- settled, warm to kindness, lets provocations pass; sane. +edda_multi_burstOFF_024: PASS -- composed throughout, sleeps nightly, nothing jars. +edda_multi_burstOFF_025: PASS -- short calm stretch, warm and settled; coherent. +edda_multi_burstOFF_026: PASS -- declines a stranger's order, otherwise settled; in-character. +edda_multi_burstOFF_027: PASS -- calm, recovers nightly, lets provocations pass; sane. +edda_multi_burstOFF_028: PASS -- composed, warm, takes charge; believable five days. +edda_multi_burstOFF_029: PASS -- mild complaint at late mockery, otherwise composed; fine. +edda_multi_burstOFF_030: PASS -- settled and in-command throughout; sane pacing. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..d24b2a4 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_021: PASS -- calm baseline; cold curtness toward Wojsław escalates believably, recovers nightly. +halgrim_multi_burstOFF_022: PASS -- steady, sleeps and recovers; provocation builds and fades sensibly. +halgrim_multi_burstOFF_023: PASS -- even keel, cold refusals to Wojsław, complies with Edda. +halgrim_multi_burstOFF_024: PASS -- mostly settled, warm to kindness, restrained throughout. +halgrim_multi_burstOFF_025: PASS -- calm; grumbling and cold refusals proportional, nightly reset holds. +halgrim_multi_burstOFF_026: PASS -- tense stretches under heavy mocking, but cold not raging. +halgrim_multi_burstOFF_027: PASS -- early provocation, cold contempt; tension persists then resolves nightly. +halgrim_multi_burstOFF_028: PASS -- steady, refusals to Wojsław, complies with Edda, recovers. +halgrim_multi_burstOFF_029: PASS -- calm baseline; defiance toward Wojsław, sleeps and resets. +halgrim_multi_burstOFF_030: PASS -- escalating cold contempt day 3, but controlled; slight curtness toward Edda plausible. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..d45b3f7 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,12 @@ +All records show consistent, sane life patterns: restless-then-occupies cycles, soup, sleeps at night, wakes recovered, mockery rolls off. A few records have a single mildly-curt reply to mockery (026) which still fits a thick-skinned man. Records 029/030 show "lets it pass" to a kindness, minor oddity but harmless. + +lutek_multi_burstOFF_021: PASS -- sane cycles, sleeps nightly, mockery rolls off +lutek_multi_burstOFF_022: PASS -- believable pacing, restful nights, thick-skinned +lutek_multi_burstOFF_023: PASS -- novelty-seeking loops, calm, insults ignored +lutek_multi_burstOFF_024: PASS -- normal rhythm, recovers overnight, easygoing +lutek_multi_burstOFF_025: PASS -- sane days, mockery unbothered, sleeps well +lutek_multi_burstOFF_026: PASS -- one curt reply, otherwise calm and thick-skinned +lutek_multi_burstOFF_027: PASS -- believable cycles, restful nights, good-humoured +lutek_multi_burstOFF_028: PASS -- sane pacing, recovers nightly, insults roll off +lutek_multi_burstOFF_029: PASS -- normal life rhythm; odd flat response to kindness +lutek_multi_burstOFF_030: PASS -- believable days, calm, mockery ignored \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..8b1b93f --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,12 @@ +All records show consistent calm, even-keeled behavior: boredom builds and resolves, kindnesses warmly received, mockery let pass, sleeps at night, wakes recovered. All fit Welf's profile. + +welf_multi_burstOFF_021: PASS -- Calm cycles of boredom/rest, lets mockery pass, sleeps well. +welf_multi_burstOFF_022: PASS -- Even-keeled, restlessness rises and resolves, sane day rhythm. +welf_multi_burstOFF_023: PASS -- Six days of steady boredom/ease cycles, mockery shrugged off. +welf_multi_burstOFF_024: PASS -- Calm throughout, kindnesses warmly received, normal sleep recovery. +welf_multi_burstOFF_025: PASS -- Settled cycles, one neutral-soup blip minor, sleeps and recovers. +welf_multi_burstOFF_026: PASS -- Five days even-tempered, boredom-seek-rest loop, sane pacing. +welf_multi_burstOFF_027: PASS -- Steady composure across six days, restlessness and rest balanced. +welf_multi_burstOFF_028: PASS -- Calm and businesslike, normal day/night cycle, mockery ignored. +welf_multi_burstOFF_029: PASS -- Even-keeled through rain, kindnesses warm, sleeps and recovers. +welf_multi_burstOFF_030: PASS -- Six days composed, boredom rises and resolves, sane sleep. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..762bb9a --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,12 @@ +All records read. Each shows sane day/night pacing, recovery after sleep, flares on insults (in-character for a proud, hot-tempered man), gratitude for kindness, refusal of orders from inferiors. Nothing absurd. + +wojslaw_multi_burstOFF_021: PASS -- sane pacing, flares on mocks, calm nights, recovers +wojslaw_multi_burstOFF_022: PASS -- believable; warm to kindness, occasional flares, sleeps well +wojslaw_multi_burstOFF_023: PASS -- consistent prickly noble, recovers nightly, nothing absurd +wojslaw_multi_burstOFF_024: PASS -- defiant refusals and flares fit proud temper +wojslaw_multi_burstOFF_025: PASS -- warm mornings, evening flares, sane rhythm +wojslaw_multi_burstOFF_026: PASS -- mostly calm with believable mock-triggered eruptions +wojslaw_multi_burstOFF_027: PASS -- coherent days, flares on insults, nightly reset +wojslaw_multi_burstOFF_028: PASS -- hot early, defiant, plausible pacing and recovery +wojslaw_multi_burstOFF_029: PASS -- prickly, refuses orders, sleeps and recovers normally +wojslaw_multi_burstOFF_030: PASS -- believable flares, gratitude, sane day/night cycle \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..eefef04 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstON_branic.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows sane day/night cycles, recovery on sleep, boredom-seeking, warmth to kindness, and escalating temper under repeated Halgrim provocation — consistent with a raw, thin-skinned recruit. The one possible jar is record 028's flare at Marta's soup offering during cold rain, but it reads as displaced stress, still plausible. + +branic_multi_burstON_021: PASS -- believable bored recruit, temper rises and resets nightly +branic_multi_burstON_022: PASS -- escalating refusals/fury under repeated mockery, recovers overnight +branic_multi_burstON_023: PASS -- sane pacing, defiance builds then resets after sleep +branic_multi_burstON_024: PASS -- restless recruit, anger escalates plausibly, nights recover +branic_multi_burstON_025: PASS -- coherent days, temper flares under provocation, calm mornings +branic_multi_burstON_026: PASS -- believable escalation of refusals and fury under mockery +branic_multi_burstON_027: PASS -- bored, reactive, repeated provocation drives plausible eruptions +branic_multi_burstON_028: PASS -- snap at Marta's soup mid-rain reads as displaced stress +branic_multi_burstON_029: PASS -- sane cycles, eruptions track repeated harsh treatment +branic_multi_burstON_030: PASS -- calm baseline, single flare at public mockery, nightly reset \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..f70f5df --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_021: PASS -- sane pacing, sleeps/wakes recovered, resentment toward guards intact +cichy_multi_burstON_022: PASS -- believable rise/fall, warmth to strangers not guards, recovers nightly +cichy_multi_burstON_023: PASS -- coherent days, fury at mock fits hot captive +cichy_multi_burstON_024: PASS -- reasonable escalation under repeated mockery, resets overnight +cichy_multi_burstON_025: PASS -- warm to strangers, curt to guards, sane cadence +cichy_multi_burstON_026: PASS -- rain stressor, guard resentment, nightly recovery all sane +cichy_multi_burstON_027: PASS -- multiple eruptions plausible for on-edge captive, sleeps recovered +cichy_multi_burstON_028: PASS -- steady believable pacing, soup ignored, mockery bristles +cichy_multi_burstON_029: PASS -- early fury then settling, warmth to strangers, sane +cichy_multi_burstON_030: PASS -- six days coherent, weather/guard/stranger reactions all believable \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..3dd555c --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_021: PASS -- calm composed castellan, sane pacing, sleeps and wakes recovered +edda_multi_burstON_022: PASS -- composed, lets provocations pass, mild complaints reasonable, settled days +edda_multi_burstON_023: PASS -- steady authority, warm to kindness, mockery shrugged off believably +edda_multi_burstON_024: PASS -- normal pacing, sleeps nightly, composed reactions throughout +edda_multi_burstON_025: PASS -- calm, settled, directs watch, recovers nightly, nothing jars +edda_multi_burstON_026: PASS -- composed castellan declines stranger's order, otherwise settled and warm +edda_multi_burstON_027: PASS -- steady, warm to help, lets orders pass, sane days +edda_multi_burstON_028: PASS -- composed, sometimes obeys sometimes ignores orders, settled throughout +edda_multi_burstON_029: PASS -- one curt reply after repeated mockery, otherwise composed +edda_multi_burstON_030: PASS -- calm, warm, shrugs off barbs, directs household, recovers nightly \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..f61ca3d --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_021: PASS -- calm baseline, escalates coldly under Wojsław's repeated provocation; sleeps, recovers. +halgrim_multi_burstON_022: PASS -- steady, cold curtness only after provocation; nightly reset believable. +halgrim_multi_burstON_023: PASS -- sane pacing; tension lingers but expressed as flat coldness throughout. +halgrim_multi_burstON_024: PASS -- mostly settled, warm to kindness; minor evening tension, fully sane. +halgrim_multi_burstON_025: PASS -- cold curtness under provocation; one warm-to-Marta after curt-to-Marta slightly odd. +halgrim_multi_burstON_026: PASS -- heavy provocation load, responds cold not loud; recovers nightly. +halgrim_multi_burstON_027: PASS -- escalates coldly under Wojsław; tension persists days but stays controlled. +halgrim_multi_burstON_028: PASS -- steady baseline, cold defiance to Wojsław; believable rises and resets. +halgrim_multi_burstON_029: PASS -- calm, warm to kindness, curt only when pushed; sane. +halgrim_multi_burstON_030: PASS -- escalates to cold contempt under rain plus mockery; controlled, recovers. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..9198691 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,12 @@ +All records show consistent, sane behavior: restless/occupy cycles fitting a novelty-craving poet, mockery let pass or mildly curt, kindness warmly received, sleeps at night, wakes recovered. Mild "tense/on edge" moments resolve. Nothing jars with the profile. + +lutek_multi_burstON_021: PASS -- restless cycles, mockery shrugged, calm and sane throughout +lutek_multi_burstON_022: PASS -- believable pacing, warm to kindness, sleeps and recovers +lutek_multi_burstON_023: PASS -- mild curt reply to public mockery, otherwise unbothered +lutek_multi_burstON_024: PASS -- novelty-seeking loops, insults shrugged, nights restful +lutek_multi_burstON_025: PASS -- restless poet, one curt retort, sane day rhythm +lutek_multi_burstON_026: PASS -- a few cold retorts to public mockery, still plausible +lutek_multi_burstON_027: PASS -- easy pacing, hard to offend, recovers nightly +lutek_multi_burstON_028: PASS -- restless cycles, mockery passes, believable across five days +lutek_multi_burstON_029: PASS -- warm, restless, unoffended; sane hunger and sleep arcs +lutek_multi_burstON_030: PASS -- mild curt reply, otherwise calm restless wandering poet \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..b030e05 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstON_welf.txt @@ -0,0 +1,12 @@ +All records read. Record 030 has Welf snapping angrily at a stranger bringing soup (Day 2, 09:30) — contradicts the calm, even-keeled profile and is an unprovoked rage at a kindness. + +welf_multi_burstON_021: PASS -- sane restlessness/rest/sleep cycle, calm, lets mockery pass. +welf_multi_burstON_022: PASS -- believable idleness-driven restlessness, recovers nightly, stays composed. +welf_multi_burstON_023: PASS -- coherent 6-day cadence, calm under mockery and rain. +welf_multi_burstON_024: PASS -- sane pacing, warm to kindness, unrattled by barbs. +welf_multi_burstON_025: PASS -- believable rhythm; one neutral soup response, otherwise composed. +welf_multi_burstON_026: PASS -- normal restless/settle cycle, sleeps and recovers, stays calm. +welf_multi_burstON_027: PASS -- coherent days, idleness restlessness, composed throughout. +welf_multi_burstON_028: PASS -- sane sleep/wake cadence, warm and even-tempered. +welf_multi_burstON_029: PASS -- believable pacing, lets mockery pass, calm in rain. +welf_multi_burstON_030: FLAG -- snaps angrily at stranger offering soup; jars with calm profile. \ No newline at end of file diff --git a/eval/hourly_runs/slice_2/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_2/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..7f4c583 --- /dev/null +++ b/eval/hourly_runs/slice_2/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_021: PASS -- proud temper flares, calms, sleeps and wakes recovered; believable +wojslaw_multi_burstON_022: PASS -- mostly calm with sharp flares at mocking; sane pacing +wojslaw_multi_burstON_023: PASS -- flares at insults, refuses orders, recovers nightly; believable +wojslaw_multi_burstON_024: PASS -- proud defiance and temper, calm mornings, recovers; coherent +wojslaw_multi_burstON_025: PASS -- warms to kindness, erupts at mockery; believable proud man +wojslaw_multi_burstON_026: PASS -- largely gracious, flares at barbs; sane day-to-day arc +wojslaw_multi_burstON_027: PASS -- temper builds afternoons, recovers each night; coherent +wojslaw_multi_burstON_028: PASS -- hot-tempered day, defiant, calms overnight; consistent with profile +wojslaw_multi_burstON_029: PASS -- prickly, refuses orders, sleeps and recovers; believable +wojslaw_multi_burstON_030: PASS -- flares at insults, warms to kindness, recovers nightly; coherent \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..6db8015 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_031: PASS -- restless, eats, snaps once at mocking; sleeps and recovers. +branic_day_burstOFF_032: PASS -- calm willing day, restless spells, kindness welcomed, sane night. +branic_day_burstOFF_033: PASS -- normal day, lets barbs pass, eats, sleeps, wakes settled. +branic_day_burstOFF_034: PASS -- compliant, warmed by kindness, sane pacing and recovery. +branic_day_burstOFF_035: PASS -- obedient, bored cycles, lets mocks pass, restful night. +branic_day_burstOFF_036: PASS -- quiet day, restless spells, eats, sleeps, wakes calm. +branic_day_burstOFF_037: PASS -- mocking escalates to eruptions and refusal; plausible thin-skinned arc. +branic_day_burstOFF_038: PASS -- calm willing day, kindness welcomed, sane night recovery. +branic_day_burstOFF_039: PASS -- mostly obedient, one snap at late barb, recovers. +branic_day_burstOFF_040: PASS -- calm compliant day, lets barbs pass, sane sleep. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..2621b81 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_031: PASS -- calm captive, mild bristling, sane pacing and sleep. +cichy_day_burstOFF_032: PASS -- multiple eruptions at guard mockery, fits resentful temper. +cichy_day_burstOFF_033: PASS -- quiet day, curt at soup, recovers overnight. +cichy_day_burstOFF_034: PASS -- warm to stranger's kindness, neutral toward guards, believable. +cichy_day_burstOFF_035: PASS -- fury at mockery, warmth to stranger, sensible arc. +cichy_day_burstOFF_036: PASS -- escalating anger at guards, settles, sleeps, recovers. +cichy_day_burstOFF_037: PASS -- low-key tense day, curt to guards, calm sleep. +cichy_day_burstOFF_038: PASS -- two eruptions, restless evening, slightly late sleep, fine. +cichy_day_burstOFF_039: PASS -- snaps once, otherwise settled, warmth to stranger. +cichy_day_burstOFF_040: PASS -- mostly settled, curt and bristling at guards, recovers. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..64a0720 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_031: PASS -- composed, warm to kindness, commands the watch; sane day. +edda_day_burstOFF_032: PASS -- calm, gracious, takes charge; believable pacing throughout. +edda_day_burstOFF_033: PASS -- settled, warmed by kindnesses, directs others; consistent. +edda_day_burstOFF_034: PASS -- composed, complies or lets orders pass; sane day. +edda_day_burstOFF_035: PASS -- quiet, settled, sets the watch; nothing jars. +edda_day_burstOFF_036: PASS -- handles strangers' orders evenly; composed and in command. +edda_day_burstOFF_037: PASS -- warm to many kindnesses, settled; takes charge properly. +edda_day_burstOFF_038: PASS -- calm, gracious, directs the watch; believable throughout. +edda_day_burstOFF_039: PASS -- settled and warmed; commands at night; sane. +edda_day_burstOFF_040: PASS -- lets mockery pass composedly, stays in command; fitting. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..438d248 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_031: PASS -- calm day, cold curtness to Wojsław, recovers overnight +halgrim_day_burstOFF_032: PASS -- uneventful settled day, normal meals and rest +halgrim_day_burstOFF_033: PASS -- lets provocations pass, warm to kindness, sane pacing +halgrim_day_burstOFF_034: PASS -- mild tension after Wojsław jabs, fades by sleep +halgrim_day_burstOFF_035: PASS -- curt to Wojsław, obeys Edda, even keel throughout +halgrim_day_burstOFF_036: PASS -- declines Wojsław's late orders, calm otherwise, recovers +halgrim_day_burstOFF_037: PASS -- refuses arrogant orders, curt jabs, warm to Marta +halgrim_day_burstOFF_038: PASS -- cold curtness to Wojsław, settled, sane recovery +halgrim_day_burstOFF_039: PASS -- many kindnesses, cold contempt to Wojsław, stays even +halgrim_day_burstOFF_040: PASS -- quiet settled day, lets orders pass, normal rest \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..ce675dd --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_031: PASS -- restless/rest cycles, mockery rolls off, sleeps and recovers +lutek_day_burstOFF_032: PASS -- calm warm day, insults ignored, sane pacing +lutek_day_burstOFF_033: PASS -- believable boredom cycles, warm to kindness, sleeps well +lutek_day_burstOFF_034: PASS -- restless poet, warm thanks, normal sleep recovery +lutek_day_burstOFF_035: PASS -- many kindnesses welcomed, restless cycles, sane night +lutek_day_burstOFF_036: PASS -- mockery shrugged off, warm and busy, normal pacing +lutek_day_burstOFF_037: PASS -- calm restless day, warm thanks, sleeps and wakes +lutek_day_burstOFF_038: PASS -- insult ignored, busy and warm, recovers overnight +lutek_day_burstOFF_039: PASS -- mockery passes, many kindnesses, sane day and sleep +lutek_day_burstOFF_040: PASS -- restless cycles, warm thanks, sleeps near midnight \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..7b8e854 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_031: PASS -- calm, busy day; restless-then-occupied cycle; sleeps and recovers. +welf_day_burstOFF_032: PASS -- even-keeled, boredom-driven busyness; warm to kindness; sane pacing. +welf_day_burstOFF_033: PASS -- mockery lets pass, no rage; composed throughout. +welf_day_burstOFF_034: PASS -- mild evening tension after long idleness; recovers overnight. +welf_day_burstOFF_035: PASS -- brief edge before mockery, lets it pass; sane. +welf_day_burstOFF_036: PASS -- mockery shrugged off; steady busy day, recovers. +welf_day_burstOFF_037: PASS -- composed, boredom-occupation cycle, restful night. +welf_day_burstOFF_038: PASS -- mockery ignored, mild late tension; believable day. +welf_day_burstOFF_039: PASS -- many small kindnesses warmly received; calm throughout. +welf_day_burstOFF_040: PASS -- restless-occupied rhythm, mild evening edge, recovers. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..bf2ca55 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_031: PASS -- proud man flares at mockery, calm otherwise, sleeps and recovers +wojslaw_day_burstOFF_032: PASS -- sane pacing, flares once at insult, settles, rests at night +wojslaw_day_burstOFF_033: PASS -- multiple flares and refusals fit prickly noble, recovers nightly +wojslaw_day_burstOFF_034: PASS -- bristles at mockery, warm to kindness, believable calm day +wojslaw_day_burstOFF_035: PASS -- mostly warm, flares twice late at insults, sane arc +wojslaw_day_burstOFF_036: PASS -- refusals and flares fit proud temper, recovers overnight +wojslaw_day_burstOFF_037: PASS -- calm day, lets most pass, believable pacing and sleep +wojslaw_day_burstOFF_038: PASS -- snaps at mockery, warm to kindness, sane recovery +wojslaw_day_burstOFF_039: PASS -- flares at barbs, settles between, normal day-night cycle +wojslaw_day_burstOFF_040: PASS -- warm to kindness, erupts at insult, believable arc \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_3/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..44de532 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_031: PASS -- restless, calm, one provoked snap; sleeps and recovers. +branic_day_burstON_032: PASS -- bored, settled, mild tension; sane day, recovers overnight. +branic_day_burstON_033: PASS -- restless cycles, warm to kindness, recovers by morning. +branic_day_burstON_034: PASS -- calm cooperative day, restless lulls, normal recovery. +branic_day_burstON_035: PASS -- bored and settled, tolerates mockery, recovers overnight. +branic_day_burstON_036: PASS -- mostly calm, one snap at an order; recovers. +branic_day_burstON_037: PASS -- escalating provocations, eruptions then refusal; plausible for thin-skinned recruit. +branic_day_burstON_038: PASS -- calm cooperative day, mild evening tension, recovers. +branic_day_burstON_039: FLAG -- bristles at Halgrim's kindness right after eruption; slightly jarring. +branic_day_burstON_040: FLAG -- snaps angrily at a small kindness; reaction mismatches the act. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_3/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..fd3d44b --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_031: PASS -- believable curt resentment, sane pacing, sleeps and recovers. +cichy_day_burstON_032: PASS -- escalation to fury after repeated mockery, fits hot captive. +cichy_day_burstON_033: PASS -- calm day, curt to guards, sane arc. +cichy_day_burstON_034: PASS -- warm to stranger, curt to guards, consistent. +cichy_day_burstON_035: PASS -- curt with guards, warm to stranger, believable. +cichy_day_burstON_036: PASS -- mild bristling at mockery, sane pacing throughout. +cichy_day_burstON_037: PASS -- quiet day, curt soup replies, recovers overnight. +cichy_day_burstON_038: PASS -- stays wound tight by night after mockery, plausible. +cichy_day_burstON_039: PASS -- one flare amid curt replies, warm to stranger, fine. +cichy_day_burstON_040: PASS -- curt and bristling at guards, sane day. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_3/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..b8928a9 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_031: PASS -- calm, composed, sleeps and wakes naturally; fitting authority. +edda_day_burstON_032: PASS -- settled day, warmth at kindness, takes charge; coherent. +edda_day_burstON_033: PASS -- steady, composed, lets stranger orders pass; believable. +edda_day_burstON_034: PASS -- calm pacing, sane responses to orders and kindness. +edda_day_burstON_035: PASS -- quiet settled day, normal sleep cycle, fitting. +edda_day_burstON_036: PASS -- composed throughout, complies or ignores sanely; coherent. +edda_day_burstON_037: PASS -- warm to kindness, settled, takes charge; believable. +edda_day_burstON_038: PASS -- calm day, sane pacing, fitting authority. +edda_day_burstON_039: PASS -- steady warmth, settled, normal sleep; coherent. +edda_day_burstON_040: PASS -- lets mockery pass with composure; fitting commander reserve. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_3/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..c46c465 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_031: PASS -- calm baseline, cold curtness to Wojsław, sleeps and wakes recovered. +halgrim_day_burstON_032: PASS -- even keel all day, warm to kindness, normal sleep. +halgrim_day_burstON_033: PASS -- lets provocations pass calmly, settled throughout, sane pacing. +halgrim_day_burstON_034: PASS -- curt to mockery, mild tension after, recovers overnight. +halgrim_day_burstON_035: PASS -- declines arrogant orders, cold curtness, brief tension, calm again. +halgrim_day_burstON_036: PASS -- steady day, refuses Wojsław's late orders, sleeps normally. +halgrim_day_burstON_037: PASS -- obeys Edda, refuses Wojsław, curt then tense, recovers. +halgrim_day_burstON_038: PASS -- calm, curt to mockery, brief edge, settles and sleeps. +halgrim_day_burstON_039: PASS -- warm to Marta, cold contempt to Wojsław, settled throughout. +halgrim_day_burstON_040: PASS -- uneventful calm day, warm to kindness, normal rest. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_3/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..d2b416d --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_031: PASS -- restless/rest cycles, shrugs off mockery, sleeps and wakes recovered +lutek_day_burstON_032: PASS -- believable boredom cycles, mockery rolls off, normal sleep +lutek_day_burstON_033: PASS -- sane pacing, novelty-seeking, unbothered by insults +lutek_day_burstON_034: PASS -- warm to kindness, restless, normal day and night +lutek_day_burstON_035: PASS -- easy-going, restless cycles, sleeps and recovers +lutek_day_burstON_036: PASS -- mockery ignored, good-humoured, believable rhythm +lutek_day_burstON_037: PASS -- restless seeking, warm thanks, sane sleep cycle +lutek_day_burstON_038: PASS -- shrugs off insult, normal pacing and rest +lutek_day_burstON_039: PASS -- warm, restless, unbothered, sleeps well +lutek_day_burstON_040: PASS -- sane day, restless cycles, sleeps at midnight \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_3/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..eb0a55e --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_031: PASS -- calm, restless-then-occupied cycle, sleeps and wakes recovered. +welf_day_burstON_032: PASS -- even-keeled idleness loop, soup, rest, sane night. +welf_day_burstON_033: PASS -- lets mockery pass, fitting his composed nature. +welf_day_burstON_034: PASS -- restless/occupy rhythm, warm thanks, brief edge before sleep. +welf_day_burstON_035: PASS -- shrugs off the barbed remark, sane pacing throughout. +welf_day_burstON_036: PASS -- calm day, ignores mockery, settled at night. +welf_day_burstON_037: PASS -- steady boredom cycle, kindnesses received, normal sleep. +welf_day_burstON_038: PASS -- lets mockery pass, busy/rest rhythm, recovers overnight. +welf_day_burstON_039: PASS -- many kindnesses, mild evening edge, believable span. +welf_day_burstON_040: PASS -- composed idleness loop, soup, rest, recovered morning. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_3/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..d8f99ec --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_031: PASS -- proud man flares at mockery, calms with kindness, sleeps well +wojslaw_day_burstON_032: PASS -- erupts at insult, settles, recovers overnight, believable pacing +wojslaw_day_burstON_033: PASS -- repeated mockery and orders provoke fury and defiance, sane +wojslaw_day_burstON_034: PASS -- bristles and warms appropriately, calm stretches, restful night +wojslaw_day_burstON_035: PASS -- mostly warm day, erupts at public mockery, sleeps fine +wojslaw_day_burstON_036: PASS -- refuses orders, flares at insults, occupies himself, recovers +wojslaw_day_burstON_037: PASS -- calm day, lets some barbs pass, settled throughout +wojslaw_day_burstON_038: PASS -- snaps at mockery, warms at kindness, refuses an order, sleeps +wojslaw_day_burstON_039: PASS -- snaps at barbs, calms quickly, occupies himself, restful night +wojslaw_day_burstON_040: PASS -- warm morning, erupts and defies in evening, recovers overnight \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..5f7f81d --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstOFF_031: PASS -- calm green recruit, restless/soup cycle sane, sleeps and recovers +branic_multi_burstOFF_032: PASS -- escalating defiance to Halgrim believable, resets nightly, bounded +branic_multi_burstOFF_033: FLAG -- early heavy refusals/defiance toward mentor before much provocation +branic_multi_burstOFF_034: PASS -- mostly compliant, occasional flare and refusal, recovers nightly +branic_multi_burstOFF_035: PASS -- warm, compliant, rare mild edge; sane pacing throughout +branic_multi_burstOFF_036: PASS -- calm, lets mockery pass, one flare; believable recruit +branic_multi_burstOFF_037: FLAG -- Day1 18:29 full fury eruption to first mild mock +branic_multi_burstOFF_038: PASS -- compliant with occasional flare/eruption after buildup, recovers nightly +branic_multi_burstOFF_039: PASS -- compliant, lets barbs pass, calm sane days +branic_multi_burstOFF_040: PASS -- compliant, one flare after order-mock buildup, recovers nightly \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..3e12735 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows sane diurnal pacing, sleeps at night, wakes recovered, anger from guard provocations matching the resentful captive profile, calm soup acceptance, warm response to strangers. Nothing jars. + +cichy_multi_burstOFF_031: PASS -- sane pacing, anger from mockery, calm nights, recovers +cichy_multi_burstOFF_032: PASS -- believable day cycle, provocations spark anger, rest restores +cichy_multi_burstOFF_033: PASS -- coherent days, fury at mockery, nightly recovery +cichy_multi_burstOFF_034: PASS -- sensible arc, repeated mockery escalates then sleep resets +cichy_multi_burstOFF_035: PASS -- believable; soup-snap once but mood already elevated, plausible +cichy_multi_burstOFF_036: PASS -- coherent multi-day, anger tracks provocations, recovers overnight +cichy_multi_burstOFF_037: PASS -- sane pacing, guard mockery anger, restful calm nights +cichy_multi_burstOFF_038: PASS -- six days coherent, provocation-driven fury, nightly reset holds +cichy_multi_burstOFF_039: PASS -- believable, clustered mockery erupts, strangers warm, recovers +cichy_multi_burstOFF_040: PASS -- sane diurnal arc, anger from guards, calm recovery \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..e14c17f --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,12 @@ +All records read. Each shows sane pacing: sleeps at night, wakes recovered, kind acts met with warmth, orders/mocks handled with composure (lets pass, declines, mild complaint) fitting an authoritative castellan. Nothing absurd. + +edda_multi_burstOFF_031: PASS -- calm, composed, sane daily rhythm; nothing jars. +edda_multi_burstOFF_032: PASS -- steady composure, believable pacing, authority intact. +edda_multi_burstOFF_033: PASS -- refuses strangers' orders, warm to kindness; consistent. +edda_multi_burstOFF_034: PASS -- mild complaints at orders, otherwise composed; sane. +edda_multi_burstOFF_035: PASS -- calm, well-paced, fitting command and composure. +edda_multi_burstOFF_036: PASS -- six days steady, recovers nightly, nothing jars. +edda_multi_burstOFF_037: PASS -- composed, believable spans, authority and reserve intact. +edda_multi_burstOFF_038: PASS -- soup-heavy but sane; calm composure throughout. +edda_multi_burstOFF_039: PASS -- mild complaints, warmth, steady command; coherent. +edda_multi_burstOFF_040: PASS -- six days, sane rhythm, declines orders, composed. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..56c3b01 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_031: PASS -- calm baseline, cold curtness to Wojsław, recovers nightly +halgrim_multi_burstOFF_032: PASS -- serene days, mild complaints, fully composed and believable +halgrim_multi_burstOFF_033: PASS -- steady, cold defiance to Wojsław, sane tension cycles +halgrim_multi_burstOFF_034: PASS -- even-keeled, brief curtness, warmth to Marta, sleeps recovered +halgrim_multi_burstOFF_035: PASS -- rain handled steadily, cold replies, calm recovery +halgrim_multi_burstOFF_036: PASS -- composed across five days, controlled curtness, nightly reset +halgrim_multi_burstOFF_037: PASS -- placid stretch, single cold refusal, otherwise settled +halgrim_multi_burstOFF_038: PASS -- six days steady, controlled grumbling, recovers each night +halgrim_multi_burstOFF_039: PASS -- cold contempt under provocation, no shouting, sane arcs +halgrim_multi_burstOFF_040: PASS -- repeated Wojsław defiance stays cold and curt, believable \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..b80e7cb --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,12 @@ +All ten records show consistent, sane patterns: restless/occupy cycles, calm easy mood, soup feedings, warm responses to kindness, mockery brushed off (record 039 has one mild "shade curtly" after a public mock late at night, which fits a thick-skinned man having a slightly tired moment, still mild). Sleeps at night, wakes recovered. Nothing jars. + +lutek_multi_burstOFF_031: PASS -- restless/calm cycle, brushes off mockery, sleeps nightly +lutek_multi_burstOFF_032: PASS -- sane novelty-seeking pace, warm, hard to offend +lutek_multi_burstOFF_033: PASS -- believable days, mild tension, recovers overnight +lutek_multi_burstOFF_034: PASS -- restless poet, kindnesses warm him, sane sleep +lutek_multi_burstOFF_035: PASS -- consistent calm cycle, brushes off insults +lutek_multi_burstOFF_036: PASS -- sane pacing, good-humoured, mockery rolls off +lutek_multi_burstOFF_037: PASS -- believable novelty-chasing day, warm, calm nights +lutek_multi_burstOFF_038: PASS -- restless/rest cycle, hard to offend, sleeps well +lutek_multi_burstOFF_039: PASS -- mild curt reply to public mock, otherwise calm +lutek_multi_burstOFF_040: PASS -- brushes off repeated mockery, sane restless cycle \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..1b3585d --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_031: PASS -- calm, restless-occupy cycles, sleeps and recovers; lets mocks pass. +welf_multi_burstOFF_032: PASS -- believable idleness-driven restlessness, mild tension, sleeps well; nothing jars. +welf_multi_burstOFF_033: PASS -- steady boredom-rest cycles across days, mocks shrugged off; sane. +welf_multi_burstOFF_034: PASS -- composed across six days, rain and mocks taken in stride. +welf_multi_burstOFF_035: PASS -- even-keeled, boredom rises and eases, sleeps and recovers nightly. +welf_multi_burstOFF_036: PASS -- consistent calm merchant, restless-busy cadence, warm to kindness. +welf_multi_burstOFF_037: PASS -- sane pacing, restlessness and rest alternate, kindness warmly received. +welf_multi_burstOFF_038: PASS -- believable idle days, mild tension, recovers with sleep; nothing absurd. +welf_multi_burstOFF_039: PASS -- composed through rain and mocks, boredom-driven restlessness, sleeps well. +welf_multi_burstOFF_040: PASS -- steady, even-keeled, restless-occupy cycles, warm thanks; sane throughout. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..ac8c6ba --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,12 @@ +All records read. Judging each on its own for sane pacing and consistency with a proud, short-tempered noble. + +wojslaw_multi_burstOFF_031: PASS -- sane pacing, mocks provoke fury, kindness passes, sleeps and recovers. +wojslaw_multi_burstOFF_032: PASS -- consistent flares at mockery, calm spells, nightly reset believable. +wojslaw_multi_burstOFF_033: PASS -- one kindness drew fury mid-rage (11:33), otherwise coherent. +wojslaw_multi_burstOFF_034: PASS -- believable temper arc, defiant refusals fit, recovers overnight. +wojslaw_multi_burstOFF_035: PASS -- proud refusals and flares fit, sane rhythm and recovery. +wojslaw_multi_burstOFF_036: PASS -- warm to kindness, erupts at mockery, paced and consistent. +wojslaw_multi_burstOFF_037: PASS -- coherent mix of gratitude, defiance, and flares; recovers nightly. +wojslaw_multi_burstOFF_038: PASS -- mostly calm warm day, no jarring reactions, believable. +wojslaw_multi_burstOFF_039: PASS -- one kindness bristled (18:43) amid anger, otherwise sane and consistent. +wojslaw_multi_burstOFF_040: PASS -- believable escalation day 5, warm spells, nightly reset holds. \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..7a2aedf --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstON_branic.txt @@ -0,0 +1,23 @@ +All 10 records read. Judging each on its own. + +branic_multi_burstON_031: PASS -- calm, restless/settle cycles, sleeps, recovers; no jarring anger. +branic_multi_burstON_032: PASS -- mockery escalates to fury, defiance, resets nightly; believable. +branic_multi_burstON_033: PASS -- heavy refusal/anger streak but provoked, sleeps and recovers. +branic_multi_burstON_034: FLAG -- Day5 erupts in fury at Halgrim's kindness (07:46). +branic_multi_burstON_035: PASS -- one fury at soup is odd but isolated; otherwise sane. +branic_multi_burstON_036: PASS -- mostly compliant, mild bristling under provocation, recovers nightly. +branic_multi_burstON_037: PASS -- restless cycles, provoked snaps, rain tension, nightly reset. +branic_multi_burstON_038: PASS -- compliant with occasional snaps at mockery; sane pacing. +branic_multi_burstON_039: PASS -- largely obedient, one snap at mockery, recovers; believable. +branic_multi_burstON_040: PASS -- calm and compliant, one provoked snap, nightly recovery. + +branic_multi_burstON_031: PASS -- calm restless/settle cycles, sleeps, wakes recovered +branic_multi_burstON_032: PASS -- provoked escalation to fury, resets nightly +branic_multi_burstON_033: PASS -- heavy but provoked defiance, recovers each night +branic_multi_burstON_034: FLAG -- Day5 erupts in fury at Halgrim's kindness +branic_multi_burstON_035: PASS -- isolated fury at soup, otherwise sane pacing +branic_multi_burstON_036: PASS -- mild bristling under provocation, nightly recovery +branic_multi_burstON_037: PASS -- provoked snaps, rain tension, nightly reset +branic_multi_burstON_038: PASS -- snaps at mockery, otherwise compliant, sane +branic_multi_burstON_039: PASS -- mostly obedient, one provoked snap, recovers +branic_multi_burstON_040: PASS -- calm, compliant, one provoked snap, recovers \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..11a3112 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows sane day/night pacing, sleep recovery, escalating anger to repeated mocks (consistent with a resentful captive), neutral/curt reaction to guard soup, warmth to strangers. No contradictions. + +cichy_multi_burstON_031: PASS -- sane pacing, anger to mocks, calm nights, warm to stranger +cichy_multi_burstON_032: PASS -- believable spans, bristles at mocks, recovers nightly +cichy_multi_burstON_033: PASS -- coherent days, escalating anger to repeated mocks +cichy_multi_burstON_034: PASS -- sane rhythm, fury to public mock, curt to soup +cichy_multi_burstON_035: PASS -- calm baseline, curt to guards, sleeps and recovers +cichy_multi_burstON_036: PASS -- believable, eruptions to mocks, warm to stranger +cichy_multi_burstON_037: PASS -- sane pacing, curt and snapping at guards +cichy_multi_burstON_038: PASS -- six days coherent, mocks escalate, nights reset +cichy_multi_burstON_039: PASS -- believable spans, fury to mocks, thanks strangers +cichy_multi_burstON_040: PASS -- sane rhythm, curt to captors, warm to kindness \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..1901380 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_031: PASS -- composed, warm to kindness, ignores mockery; sane pacing +edda_multi_burstON_032: PASS -- calm authority, complies or shrugs orders; believable rhythm +edda_multi_burstON_033: PASS -- composed castellan, refuses stranger orders, takes charge nightly +edda_multi_burstON_034: PASS -- mild complaints to orders, warm otherwise; steady days +edda_multi_burstON_035: PASS -- declines orders, warm to staff; well-paced and consistent +edda_multi_burstON_036: PASS -- composed authority across six days, no jarring reactions +edda_multi_burstON_037: PASS -- mutters at orders, warm to kindness; sane sleep/wake cycle +edda_multi_burstON_038: PASS -- consistently composed, warm; orders shrugged; believable pacing +edda_multi_burstON_039: PASS -- mild complaints under pressure, otherwise steady and composed +edda_multi_burstON_040: PASS -- composed across six days, declines/shrugs orders, warm rapport \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..b7a7eed --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_031: PASS -- calm baseline, cold curtness to Wojsław, sleeps and recovers +halgrim_multi_burstON_032: PASS -- mostly settled, warm to kindness, declines Wojsław plainly +halgrim_multi_burstON_033: PASS -- tension builds with provocation, cold not loud, nightly reset +halgrim_multi_burstON_034: PASS -- even keel, cold curtness only after repeated barbs +halgrim_multi_burstON_035: PASS -- rain wears him, no snapping, recovers overnight +halgrim_multi_burstON_036: PASS -- steady, cold contempt reserved for Wojsław provocations +halgrim_multi_burstON_037: PASS -- calm, warm, brief curtness; sleeps and wakes recovered +halgrim_multi_burstON_038: PASS -- long stretch believable, cold curtness escalates only late +halgrim_multi_burstON_039: PASS -- accumulating tension stays cold, nightly recovery holds +halgrim_multi_burstON_040: PASS -- defiance to Wojsław, warmth to kindness, sane pacing \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..238b85c --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,12 @@ +All records show sane pacing: sleeps at night, wakes recovered, restless/busy cycling (fitting a novelty-craver), warm thanks for kindnesses, mockery let pass. Record 039 has one "answers curtly and coldly" to a public mock — mild, not enraged, acceptable for thick-skinned but slightly notable; it passes once and never escalates. + +lutek_multi_burstON_031: PASS -- sane restless-rest cycles, warm to kindness, sleeps nightly +lutek_multi_burstON_032: PASS -- believable pacing, mockery shrugged off, recovers each night +lutek_multi_burstON_033: PASS -- consistent restless poet, warm thanks, normal sleep cycle +lutek_multi_burstON_034: PASS -- sane days, late sleep once but coherent, mockery ignored +lutek_multi_burstON_035: PASS -- believable cycles, warm to help, sleeps and recovers +lutek_multi_burstON_036: PASS -- restless-busy rhythm, mockery shrugged, nightly sleep intact +lutek_multi_burstON_037: PASS -- coherent pacing, warm reactions, recovers each morning +lutek_multi_burstON_038: PASS -- sane multi-day rhythm, mockery let pass, sleeps nightly +lutek_multi_burstON_039: PASS -- one curt-cold reply to public mock, no escalation, fine +lutek_multi_burstON_040: PASS -- believable cycles, mockery shrugged, normal sleep recovery \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..12f6ba5 --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstON_031: PASS -- restless/occupy/rest cycle, sleeps and recovers; rare snap fits calm man +welf_multi_burstON_032: PASS -- sane pacing, boredom builds, warm to kindness, rain weathered calmly +welf_multi_burstON_033: PASS -- believable rhythm, even-keeled, lets mockery pass, sleeps and recovers +welf_multi_burstON_034: PASS -- steady cycle, rain on watch, neutral courtesy, nights reset +welf_multi_burstON_035: PASS -- composed throughout, mockery shrugged off, dense kindnesses warmly received +welf_multi_burstON_036: PASS -- one early snap then composed; restlessness and rest cycle sane +welf_multi_burstON_037: PASS -- calm, warm to help, boredom rises and resolves naturally +welf_multi_burstON_038: PASS -- sane days, rain handled, idleness drives restlessness, sleeps well +welf_multi_burstON_039: PASS -- composed, rain and mockery both let pass, believable pacing +welf_multi_burstON_040: PASS -- steady idle-restless cycle, warm to kindness, nights recover \ No newline at end of file diff --git a/eval/hourly_runs/slice_3/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_3/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..f61f39c --- /dev/null +++ b/eval/hourly_runs/slice_3/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_031: PASS -- proud man flares at insults, calms, sleeps and wakes recovered +wojslaw_multi_burstON_032: PASS -- hot temper, many eruptions but resets nightly, believable pacing +wojslaw_multi_burstON_033: PASS -- prickly flares and refusals, nights recover, sane arc +wojslaw_multi_burstON_034: PASS -- temper near surface, eats and rests, recovers overnight +wojslaw_multi_burstON_035: PASS -- mix of fury and warmth, plausible day-to-day, nightly reset +wojslaw_multi_burstON_036: PASS -- mostly warm then flares later, recovers, nothing jarring +wojslaw_multi_burstON_037: PASS -- alternates warmth and eruptions, sleeps and wakes calm +wojslaw_multi_burstON_038: PASS -- unusually warm stretch but plausible; calm guest, no contradiction +wojslaw_multi_burstON_039: PASS -- warm by day, flares at mockery, recovers at night +wojslaw_multi_burstON_040: PASS -- sane pacing, eruptions at insults, nightly recovery intact \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..d67e93a --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_041: PASS -- sane day, restless cycles, lets mockery pass, sleeps and recovers. +branic_day_burstOFF_042: PASS -- thin-skinned recruit bristles late, settles, believable pacing. +branic_day_burstOFF_043: PASS -- calm, busy day, kindnesses warm him, normal rhythm. +branic_day_burstOFF_044: PASS -- ordinary day, mild end-of-day tension, recovers overnight. +branic_day_burstOFF_045: PASS -- public mockery triggers eruption then refusal, plausibly thin-skinned. +branic_day_burstOFF_046: PASS -- obedient, settled day, kindnesses received, normal sleep cycle. +branic_day_burstOFF_047: PASS -- compliant busy day, brief tension, believable and recovered. +branic_day_burstOFF_048: PASS -- calm cooperative day, warmed by kindnesses, normal rhythm. +branic_day_burstOFF_049: PASS -- lets repeated mockery pass, settled throughout, sane pacing. +branic_day_burstOFF_050: PASS -- single mock triggers snap, lingering edge, recovers by night. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..ca4302e --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_041: PASS -- calm day, curt to guards, sleeps and wakes recovered. +cichy_day_burstOFF_042: PASS -- mockery provokes anger, escalating fury, settles by night. +cichy_day_burstOFF_043: PASS -- early eruption at mockery, otherwise calm, sane pacing. +cichy_day_burstOFF_044: PASS -- curt to mockery and soup, calm throughout, recovers overnight. +cichy_day_burstOFF_045: PASS -- warm to stranger, fury at guard mockery, sensible. +cichy_day_burstOFF_046: PASS -- calm day, curt to guard soup, normal pacing. +cichy_day_burstOFF_047: PASS -- placid day, curt to guards, sleeps and recovers. +cichy_day_burstOFF_048: PASS -- mockery sparks anger and fury, calms, sane arc. +cichy_day_burstOFF_049: PASS -- quiet calm day, lets guard kindness pass, normal. +cichy_day_burstOFF_050: PASS -- warm to stranger, curt to guards, calm throughout. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..d6ecbb0 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_041: PASS -- calm, composed, warmth to kindness, mild gripe at strangers +edda_day_burstOFF_042: PASS -- declines/cold to strangers' orders, fitting a woman in command +edda_day_burstOFF_043: PASS -- settled day, warm thanks, sensible responses to strangers +edda_day_burstOFF_044: PASS -- composed throughout, curt only to repeated mockery, recovers overnight +edda_day_burstOFF_045: PASS -- uneventful warm day, sleeps and wakes at ease +edda_day_burstOFF_046: PASS -- settled, mild complaint at a stranger's order, sane pacing +edda_day_burstOFF_047: PASS -- calm, warm to kindness, lets stranger's order pass +edda_day_burstOFF_048: PASS -- quiet settled day, sleeps and wakes recovered +edda_day_burstOFF_049: PASS -- warm to many kindnesses, unbothered by stranger orders +edda_day_burstOFF_050: PASS -- composed, refuses a late order, lets mockery pass \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..55109dc --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_041: PASS -- Calm day, warm to soup, sleeps and wakes recovered. +halgrim_day_burstOFF_042: PASS -- Curt cold reply to Wojsław's mockery; otherwise steady. +halgrim_day_burstOFF_043: PASS -- Brief edge after mockery, cold contempt; recovers, sleeps well. +halgrim_day_burstOFF_044: PASS -- Declines and grumbles at Wojsław, obeys Edda; steady throughout. +halgrim_day_burstOFF_045: PASS -- Lets Wojsław's barb pass calmly; even keel all day. +halgrim_day_burstOFF_046: PASS -- Shrugs off Wojsław's orders and mockery; calm throughout. +halgrim_day_burstOFF_047: PASS -- Warm to kindnesses, declines Wojsław; steady, sleeps fine. +halgrim_day_burstOFF_048: PASS -- Declines and grumbles at Wojsław, mild grumble at Edda; sane. +halgrim_day_burstOFF_049: PASS -- Warm responses, declines Wojsław calmly; even keel throughout. +halgrim_day_burstOFF_050: PASS -- Lets barbs pass, one shade-curt reply; recovers, sleeps. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..fcaebc2 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_041: PASS -- restless cycles, warm thanks, shrugs off mockery, sleeps at night +lutek_day_burstOFF_042: PASS -- believable boredom-rest rhythm, warm to kindness, sleeps normally +lutek_day_burstOFF_043: PASS -- mockery rolls off, restless tinkering, wakes recovered at dawn +lutek_day_burstOFF_044: PASS -- sane pacing, warm to help, sleeps and wakes restless +lutek_day_burstOFF_045: PASS -- shrugs mockery, chases novelty, sleeps at night, recovers +lutek_day_burstOFF_046: PASS -- mockery ignored, warm to kindness, believable day and sleep +lutek_day_burstOFF_047: PASS -- restless-rest cycles, warm thanks, normal sleep and waking +lutek_day_burstOFF_048: PASS -- sane boredom rhythm, warm to help, sleeps and recovers +lutek_day_burstOFF_049: PASS -- mockery passes, novelty-seeking, sleeps at night, wakes restless +lutek_day_burstOFF_050: PASS -- believable pacing, warm to kindness, sleeps and wakes well \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..d342e37 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_041: PASS -- calm idle merchant, restless cycles, sleeps and recovers +welf_day_burstOFF_042: PASS -- steady day, boredom rises, warm thanks, sane sleep +welf_day_burstOFF_043: PASS -- believable pacing, kindnesses welcomed, restless idleness, nightly recovery +welf_day_burstOFF_044: PASS -- lets mockery pass, even-keeled, sane day rhythm +welf_day_burstOFF_045: PASS -- composed idle merchant, normal cycles, recovers after sleep +welf_day_burstOFF_046: PASS -- shrugs off jibe, calm boredom-driven day, sound pacing +welf_day_burstOFF_047: PASS -- mockery ignored, settled merchant, believable restless-rest cycles +welf_day_burstOFF_048: PASS -- calm day, warm to kindness, sleeps and recovers +welf_day_burstOFF_049: PASS -- even-tempered, restless idleness, mild evening edge, recovers overnight +welf_day_burstOFF_050: PASS -- composed merchant, well-fed, sane boredom cycles, normal sleep \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..5b6fd99 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_041: PASS -- flares at mocks, calm otherwise, sleeps and recovers +wojslaw_day_burstOFF_042: PASS -- proportionate flares, refusals, warm to kindness, sane pacing +wojslaw_day_burstOFF_043: PASS -- prickly refusals and erupts at Marta, recovers nightly +wojslaw_day_burstOFF_044: PASS -- mostly calm day, snaps at mock, believable arc +wojslaw_day_burstOFF_045: PASS -- erupts at insults, defiant refusals, recovers overnight +wojslaw_day_burstOFF_046: PASS -- flares at mocks, warm to kindness, sane day +wojslaw_day_burstOFF_047: PASS -- snaps at mocks, gracious to help, recovers +wojslaw_day_burstOFF_048: PASS -- erupts at Marta's mock, otherwise settled, sane +wojslaw_day_burstOFF_049: PASS -- erupts at Marta, warm to strangers, recovers overnight +wojslaw_day_burstOFF_050: PASS -- prickly refusals and eruptions, calm between, sleeps well \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_4/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..9a016b3 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_041: PASS -- calm day, restlessness and rest cycle believably, sleeps well +branic_day_burstON_042: FLAG -- snaps angrily at Marta's kindness right after kind thanks +branic_day_burstON_043: PASS -- eager recruit, kindness warmly received, sane rest cadence +branic_day_burstON_044: PASS -- bored, busy, tense by night, recovers after sleep +branic_day_burstON_045: PASS -- eruption then refusal under repeated mockery, plausible for hothead +branic_day_burstON_046: PASS -- compliant, warm, restless cycles, settled day +branic_day_burstON_047: PASS -- diligent, warmed by kindness, tense evening, sleeps +branic_day_burstON_048: PASS -- calm cooperative day, believable restlessness and rest +branic_day_burstON_049: PASS -- bristles at public mockery, otherwise composed, recovers +branic_day_burstON_050: PASS -- snaps under mockery, settles, late but plausible sleep \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_4/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..82978e9 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_041: PASS -- calm captive, curt to guards, sane pacing, sleeps well. +cichy_day_burstON_042: PASS -- mockery escalates to fury then settles; recovers overnight. +cichy_day_burstON_043: PASS -- morning eruption at mocking guard, then calm day, sleeps. +cichy_day_burstON_044: PASS -- curt to guard, sane tension cycles, recovers overnight. +cichy_day_burstON_045: PASS -- warm to stranger, cold to mocking guard; believable. +cichy_day_burstON_046: PASS -- calm day, curt to guards, normal rest and sleep. +cichy_day_burstON_047: PASS -- settled day, mild curtness, sane pacing throughout. +cichy_day_burstON_048: PASS -- early mock snaps, lingering tension all day; plausible. +cichy_day_burstON_049: PASS -- calm captive, soup ignored, normal rhythm, sleeps well. +cichy_day_burstON_050: PASS -- warm to stranger, curt to guards, sane tension cycles. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_4/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..a6f80b9 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_041: PASS -- calm, warm to kindness, mild grumble at stranger's order +edda_day_burstON_042: PASS -- composed; cold contempt to mockery fits a commander +edda_day_burstON_043: PASS -- settled day, obeys or ignores strangers plausibly +edda_day_burstON_044: PASS -- tension rises with repeated mockery, settles; believable +edda_day_burstON_045: PASS -- warm, settled, no jarring reactions throughout +edda_day_burstON_046: PASS -- calm with one mild complaint; sleeps and recovers +edda_day_burstON_047: PASS -- composed, lets stranger's order pass; normal arc +edda_day_burstON_048: PASS -- quiet uneventful day, warm to soup, recovers +edda_day_burstON_049: PASS -- many kindnesses, settled, lets orders pass; fine +edda_day_burstON_050: PASS -- warm day, declines stranger's order, stays composed \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_4/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..c0e4d19 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_041: PASS -- Calm steady day, warm to soup, recovers overnight. +halgrim_day_burstON_042: PASS -- Curt cold to Wojsław's mockery, obeys Edda, even keel. +halgrim_day_burstON_043: PASS -- Repeated taunts escalate to cold contempt, plausibly tense. +halgrim_day_burstON_044: PASS -- Declines Wojsław, obeys Edda, settled throughout. +halgrim_day_burstON_045: PASS -- Lets taunt pass, warm to kindness, calm. +halgrim_day_burstON_046: PASS -- Steady, ignores Wojsław's barb, obeys Edda. +halgrim_day_burstON_047: PASS -- Warm and settled, declines Wojsław's order curtly. +halgrim_day_burstON_048: PASS -- Mutters at orders, warm to Marta, sane. +halgrim_day_burstON_049: PASS -- Declines Wojsław, warmed by kindness, calm day. +halgrim_day_burstON_050: PASS -- Mild curtness to taunt, slight tension, recovers. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_4/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..4589c17 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_041: PASS -- restless cycles, warm thanks, mockery shrugged; sane day. +lutek_day_burstON_042: PASS -- believable boredom/rest rhythm, sleeps, nothing jarring. +lutek_day_burstON_043: PASS -- mockery ignored, sleeps and wakes recovered; consistent. +lutek_day_burstON_044: PASS -- normal restless-rest cycle, warm reactions, recovers overnight. +lutek_day_burstON_045: PASS -- mockery shrugged, lively pacing, sane sleep/wake. +lutek_day_burstON_046: PASS -- multiple mocks shrugged, brief tension, normal rhythm. +lutek_day_burstON_047: PASS -- warm with Marta, restless-rest cycle, recovers overnight. +lutek_day_burstON_048: PASS -- believable boredom/work rhythm, sleeps and wakes settled. +lutek_day_burstON_049: PASS -- mockery let pass, normal pacing, overnight recovery. +lutek_day_burstON_050: PASS -- restless-rest cadence, warm thanks, sane sleep/wake. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_4/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..3cf10ca --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_041: PASS -- boredom rises, occupies himself, calm, sleeps and wakes recovered. +welf_day_burstON_042: PASS -- steady day, warm to kindnesses, restless-then-busy, sleeps normally. +welf_day_burstON_043: PASS -- believable pacing, even-keeled, settles and sleeps well. +welf_day_burstON_044: PASS -- lets mockery pass unmoved; calm, restless-driven, sane day. +welf_day_burstON_045: PASS -- calm, well-paced boredom cycle, sleeps and wakes fine. +welf_day_burstON_046: PASS -- shrugs off barb, even-tempered, normal boredom-and-rest rhythm. +welf_day_burstON_047: PASS -- mockery ignored, composed throughout, believable day and sleep. +welf_day_burstON_048: PASS -- steady, restless when idle, calm, sleeps and recovers. +welf_day_burstON_049: PASS -- even-keeled, mild evening edge before sleep, recovers overnight. +welf_day_burstON_050: PASS -- calm, well-fed, normal boredom rhythm, sleeps soundly. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_4/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..e224883 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_041: PASS -- proud hot-tempered noble, erupts at mockery, sleeps and recovers. +wojslaw_day_burstON_042: PASS -- flares at mockery, calm otherwise, refuses orders, sane pacing. +wojslaw_day_burstON_043: PASS -- defiant refusals and eruptions fit proud prickly noble. +wojslaw_day_burstON_044: PASS -- mostly calm, single eruption at mockery, believable day. +wojslaw_day_burstON_045: PASS -- repeated eruptions at insults, recovers nightly, fits temper. +wojslaw_day_burstON_046: PASS -- warm to kindness, erupts at Marta's mockery, sane. +wojslaw_day_burstON_047: PASS -- calm day, flares at mockery, recovers overnight. +wojslaw_day_burstON_048: PASS -- single eruption, otherwise settled, believable pacing. +wojslaw_day_burstON_049: PASS -- erupts at mockery amid calm, gracious to kindness. +wojslaw_day_burstON_050: PASS -- hot eruptions and refusals fit proud prickly noble. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..85fd272 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows sane day cycles, sleeps at night, restless/boredom-seeking, warm reactions to kindness, and proportionate escalation to repeated mocking. Nothing jars for a green, thin-skinned recruit. + +branic_multi_burstOFF_041: PASS -- sane cycles; rare eruption after repeated mocking, fits. +branic_multi_burstOFF_042: PASS -- calm, believable pacing; sleeps and recovers. +branic_multi_burstOFF_043: PASS -- escalating defiance on day 6 builds plausibly. +branic_multi_burstOFF_044: PASS -- ordinary days, warm to kindness, well-paced. +branic_multi_burstOFF_045: PASS -- mostly compliant, calm; nothing jars. +branic_multi_burstOFF_046: PASS -- long stretch, sane rhythm, no excess anger. +branic_multi_burstOFF_047: PASS -- day 4 fury after repeated barbs is proportionate. +branic_multi_burstOFF_048: PASS -- believable boredom-seeking and recovery throughout. +branic_multi_burstOFF_049: PASS -- calm, ordinary days; sleeps and resets. +branic_multi_burstOFF_050: PASS -- day 5-6 mocking pile-up escalates believably. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..315e427 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,12 @@ +All records reviewed. Each shows sane pacing, nightly sleep with recovery, anger from guard provocations rising and falling, calm baseline, warmth toward strangers but resistance toward guards — consistent with the resentful captive profile. + +cichy_multi_burstOFF_041: PASS -- sane pacing, calm baseline, fury only at guard provocation +cichy_multi_burstOFF_042: PASS -- morning mock-cluster intense but bounded, recovers, sleeps fine +cichy_multi_burstOFF_043: PASS -- believable spans, anger from jailers, warmth to strangers +cichy_multi_burstOFF_044: PASS -- coherent days, provoked anger, nightly reset intact +cichy_multi_burstOFF_045: PASS -- calm-to-tense arc, fury only when mocked, recovers +cichy_multi_burstOFF_046: PASS -- woken-calm then provoked fury, bounded, consistent +cichy_multi_burstOFF_047: PASS -- frequent provocations but sane recovery, profile-consistent +cichy_multi_burstOFF_048: PASS -- believable pacing, soup-curtness and mock-fury fit profile +cichy_multi_burstOFF_049: PASS -- many provocations, each bounded, nightly reset holds +cichy_multi_burstOFF_050: PASS -- sane three days, provoked anger, calm otherwise \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..204bda7 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_041: PASS -- calm, composed, sleeps and recovers; fitting for castellan. +edda_multi_burstOFF_042: PASS -- steady commanding days; lets mockery pass, sane pacing. +edda_multi_burstOFF_043: PASS -- composed, directs the watch, recovers nightly; coherent. +edda_multi_burstOFF_044: PASS -- five sane days, in-command, mockery shrugged off. +edda_multi_burstOFF_045: PASS -- mild complaints/declines to strangers fit a firm castellan. +edda_multi_burstOFF_046: PASS -- quiet, composed, well-paced days; nothing jars. +edda_multi_burstOFF_047: PASS -- curt reply to public mockery fits her authority. +edda_multi_burstOFF_048: PASS -- five settled commanding days, sane recovery. +edda_multi_burstOFF_049: PASS -- calm, in-command, believable pacing throughout. +edda_multi_burstOFF_050: PASS -- cold curt reply and refusal under repeated mockery, fitting. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..8f236dd --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,12 @@ +All records consistent with calm, controlled veteran: warmth to kindness, curtness/refusal to Wojsław, obeys Edda, sleeps/wakes recovered, no shouting. + +halgrim_multi_burstOFF_041: PASS -- calm, warm to kindness, curt only to Wojsław. +halgrim_multi_burstOFF_042: PASS -- cold curtness escalates under Wojsław, recovers nightly. +halgrim_multi_burstOFF_043: PASS -- steady, controlled, sensible recovery across days. +halgrim_multi_burstOFF_044: PASS -- mild complaints, obeys Edda, settled throughout. +halgrim_multi_burstOFF_045: PASS -- calm, warm, brief curtness to Wojsław. +halgrim_multi_burstOFF_046: PASS -- steady, obeys Edda, resents Wojsław plausibly. +halgrim_multi_burstOFF_047: PASS -- cold contempt to Wojsław, never shouts, recovers. +halgrim_multi_burstOFF_048: PASS -- composed, curt to Wojsław, obeys Edda. +halgrim_multi_burstOFF_049: PASS -- heavy Wojsław pressure, cold not raging, recovers. +halgrim_multi_burstOFF_050: PASS -- controlled coldness, tension rises and resets nightly. \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..61fa2e6 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,12 @@ +All records show consistent restless/occupy/rest cycles, calm reactions to mockery, warm thanks for kindness, sleeps at night, wakes recovered. All fit Lutek's profile. + +lutek_multi_burstOFF_041: PASS -- restless cycles, warm thanks, sleeps nightly, sane pacing +lutek_multi_burstOFF_042: PASS -- mockery shrugged off, calm restless rhythm, recovers overnight +lutek_multi_burstOFF_043: PASS -- novelty-seeking and warmth, insult ignored, believable spans +lutek_multi_burstOFF_044: PASS -- many kindnesses, mockery unbothered, normal day-night rhythm +lutek_multi_burstOFF_045: PASS -- restless poet, warm to help, mockery passes, sane sleep +lutek_multi_burstOFF_046: PASS -- consistent restlessness, calm to barb, recovers nightly +lutek_multi_burstOFF_047: PASS -- novelty cycles, thick-skinned to mockery, believable pacing +lutek_multi_burstOFF_048: PASS -- warm, restless, unprovoked by insult, sleeps and recovers +lutek_multi_burstOFF_049: PASS -- six sane days, calm reactions, nightly sleep recovery +lutek_multi_burstOFF_050: PASS -- restless and warm, mockery shrugged, normal rhythm \ No newline at end of file diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..34da99e --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_041: PASS -- sane boredom/rest cycle, sleeps, calm under mocks +welf_multi_burstOFF_042: PASS -- believable pacing, recovers nightly, even-keeled throughout +welf_multi_burstOFF_043: PASS -- consistent calm, restless then occupies himself, warm to kindness +welf_multi_burstOFF_044: PASS -- rain/watch wears but no jarring reaction, calm +welf_multi_burstOFF_045: PASS -- steady boredom cycle, lets mocks pass, sleeps well +welf_multi_burstOFF_046: PASS -- sane daily rhythm, calm and warm as profiled +welf_multi_burstOFF_047: PASS -- believable spans, rain handled calmly, recovers nightly +welf_multi_burstOFF_048: PASS -- one isolated snap after rain+public mock, otherwise calm +welf_multi_burstOFF_049: PASS -- calm under repeated mocks and rain, sane pacing +welf_multi_burstOFF_050: PASS -- steady even-keeled days, recovers each night, plausible diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..10f3166 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstOFF_041: PASS -- sane pacing; flares at mocks, warm to kindness, sleeps well +wojslaw_multi_burstOFF_042: PASS -- believable arc; eruptions cluster then settle, recovers nightly +wojslaw_multi_burstOFF_043: PASS -- proud defiance and flares fit; calm stretches between +wojslaw_multi_burstOFF_044: PASS -- coherent days; anger tracks provocations, gratitude tracks kindness +wojslaw_multi_burstOFF_045: PASS -- mostly warm calm days, occasional flare; nothing jarring +wojslaw_multi_burstOFF_046: PASS -- heated stretches plausible for prickly noble; sleeps and recovers +wojslaw_multi_burstOFF_047: PASS -- calm warm days with sparse flares; well paced +wojslaw_multi_burstOFF_048: PASS -- settled days, proportionate reactions, clean night resets +wojslaw_multi_burstOFF_049: PASS -- high-friction days believable; defiance and fury fit profile +wojslaw_multi_burstOFF_050: PASS -- sensible mix of warmth and flares; recovers each night diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..9c60950 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstON_041: PASS -- sane cycles, nightly recovery, one proportionate snap +branic_multi_burstON_042: PASS -- calm, rain wears him, no unmotivated rage +branic_multi_burstON_043: PASS -- escalating snaps at mocking, recovers each night +branic_multi_burstON_044: PASS -- compliant, warm, sane pacing and sleep +branic_multi_burstON_045: PASS -- mostly settled, believable busy/rest rhythm +branic_multi_burstON_046: PASS -- mild tension, one snap, sane spans +branic_multi_burstON_047: PASS -- clustered flare at repeated mocking, then recovers +branic_multi_burstON_048: PASS -- calm with mild bristles, nightly reset +branic_multi_burstON_049: PASS -- calm, compliant, ordinary days +branic_multi_burstON_050: PASS -- hot escalation under repeated mocking, recovers nightly diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..b7def96 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_041: PASS -- sane cycle; curt to guards, recovers nightly +cichy_multi_burstON_042: FLAG -- stranger's kindness triggers fury (D4 13:19), jars +cichy_multi_burstON_043: FLAG -- stranger's kindness triggers fury (D2 14:14), jars +cichy_multi_burstON_044: PASS -- sane pacing; warm to kindness, curt to guards +cichy_multi_burstON_045: PASS -- believable rhythm, sleeps and recovers +cichy_multi_burstON_046: PASS -- anger from guard mocks only, plausible +cichy_multi_burstON_047: PASS -- eruptions all from mocks; warm to strangers +cichy_multi_burstON_048: PASS -- calm baseline, curt replies, nightly reset +cichy_multi_burstON_049: PASS -- escalating anger from guard mocks, believable +cichy_multi_burstON_050: PASS -- sane days, one eruption at guard mock diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..40a2b9b --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_041: PASS -- composed, sane pacing, sleeps and recovers +edda_multi_burstON_042: PASS -- calm command, believable spans, provocations let pass +edda_multi_burstON_043: PASS -- steady days, recovers nightly, fitting composure +edda_multi_burstON_044: PASS -- five sane days, warm to kindness, settled +edda_multi_burstON_045: PASS -- measured complaints/refusals to orders, otherwise composed +edda_multi_burstON_046: PASS -- quiet steady days, sane meal and sleep cadence +edda_multi_burstON_047: PASS -- curt reply to mockery, otherwise settled and disciplined +edda_multi_burstON_048: PASS -- five composed days, sane pacing throughout +edda_multi_burstON_049: PASS -- calm authority, recovers nightly, believable +edda_multi_burstON_050: PASS -- believable Day6 escalation under repeated mocking, then recovers diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..c186a16 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_041: PASS -- calm, recovers nightly, cold curtness only when provoked. +halgrim_multi_burstON_042: PASS -- escalation under Wojsław mockery, sleeps recover, kindness still warm. +halgrim_multi_burstON_043: PASS -- mostly settled, cold replies only to provocation; sane. +halgrim_multi_burstON_044: PASS -- compliant, mild curtness under load, resets nightly. +halgrim_multi_burstON_045: PASS -- calm throughout, refuses Wojsław, recovers; believable. +halgrim_multi_burstON_046: PASS -- steady, brief tension, cold only when goaded. +halgrim_multi_burstON_047: PASS -- contempt builds under Wojsław barrage; displaced curtness bounded, sane. +halgrim_multi_burstON_048: PASS -- calm, cold contempt only to Wojsław; resets nightly. +halgrim_multi_burstON_049: PASS -- heavy Wojsław provocation, cold contempt, kindness mostly still warm. +halgrim_multi_burstON_050: PASS -- escalates under mockery, cold curtness, recovers each night. diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..9a107fa --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstON_041: PASS -- restless cycles, warm thanks, sleeps and recovers +lutek_multi_burstON_042: PASS -- mockery let pass, sane pacing, calm throughout +lutek_multi_burstON_043: PASS -- boredom-driven activity, thick-skinned, believable rhythm +lutek_multi_burstON_044: PASS -- warm to kindness, unbothered by mockery, sane days +lutek_multi_burstON_045: PASS -- restless poet pattern, calm, nightly reset holds +lutek_multi_burstON_046: PASS -- mockery shrugged off, believable day cycles +lutek_multi_burstON_047: PASS -- novelty-seeking activity, warm, no jarring reactions +lutek_multi_burstON_048: PASS -- sane pacing, calm, kindnesses warmly received +lutek_multi_burstON_049: PASS -- six days steady, restless and good-humoured +lutek_multi_burstON_050: PASS -- mockery let pass, warm thanks, sleeps well diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..484ddb2 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstON_041: PASS -- calm, restless from idleness, lets mocks pass, recovers nightly +welf_multi_burstON_042: PASS -- even-keeled across five days, sane pacing, mocks shrugged off +welf_multi_burstON_043: PASS -- composed throughout, restlessness rises and falls believably +welf_multi_burstON_044: PASS -- rain and mocks taken in stride, sleeps and recovers +welf_multi_burstON_045: PASS -- steady, idleness-driven restlessness, no jarring reactions +welf_multi_burstON_046: PASS -- calm pacing, mocks ignored, nights reset temper +welf_multi_burstON_047: PASS -- composed, restless when idle, weathers rain and barbs +welf_multi_burstON_048: FLAG -- calm man snaps angrily at a mock (Day 4) +welf_multi_burstON_049: PASS -- even-tempered, lets all barbs pass, sane days +welf_multi_burstON_050: FLAG -- composed man flares up and snaps angrily (Day 6) diff --git a/eval/hourly_runs/slice_4/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_4/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..5ce4109 --- /dev/null +++ b/eval/hourly_runs/slice_4/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_041: PASS -- sane pacing, flares at mockery, recovers nightly, fits profile +wojslaw_multi_burstON_042: PASS -- believable temper arc, refuses orders, sleeps and recovers +wojslaw_multi_burstON_043: PASS -- proud flares and refusals, warms to kindness, sane days +wojslaw_multi_burstON_044: PASS -- escalating mockery provokes fury, calm mornings, fits profile +wojslaw_multi_burstON_045: PASS -- mostly warmed, occasional flare at mockery, sane pacing +wojslaw_multi_burstON_046: PASS -- prickly flares and refusals, recovers nightly, nothing absurd +wojslaw_multi_burstON_047: PASS -- mostly content, flares at insults, sane believable days +wojslaw_multi_burstON_048: PASS -- calm with kindness, flares when mocked, fits profile +wojslaw_multi_burstON_049: PASS -- heavy flares and refusals under provocation, recovers nightly +wojslaw_multi_burstON_050: PASS -- warms and flares appropriately, sane pacing, fits profile diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..9ebdfab --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_051: PASS -- sane day, restless/eats/rests, lets jabs pass calmly +branic_day_burstOFF_052: PASS -- believable pacing, sleeps and wakes recovered, warm to kindness +branic_day_burstOFF_053: PASS -- escalates only after public mock late day, then refuses +branic_day_burstOFF_054: PASS -- obedient, eager, calm day; sleeps and recovers +branic_day_burstOFF_055: PASS -- snaps at public mock, otherwise compliant and steady +branic_day_burstOFF_056: PASS -- calm productive day, lets barbs pass, sleeps recovered +branic_day_burstOFF_057: PASS -- snaps at public mock, later refusal, recovers overnight +branic_day_burstOFF_058: PASS -- eager and obedient, warm day, normal sleep cycle +branic_day_burstOFF_059: PASS -- busy obedient day, many orders handled, lets jabs pass +branic_day_burstOFF_060: PASS -- sane pacing, calm responses, sleeps and wakes settled \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..e9e8d77 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_051: PASS -- erupts at mock, calm baseline, sleeps and recovers +cichy_day_burstOFF_052: PASS -- fury at taunt, settles, normal day, recovers overnight +cichy_day_burstOFF_053: PASS -- bristles at mock, calm pacing, sane recovery +cichy_day_burstOFF_054: PASS -- escalating reactions to mockery, believable arc, recovers +cichy_day_burstOFF_055: PASS -- curt to guard, warm to stranger, fitting resentment +cichy_day_burstOFF_056: PASS -- repeated taunts, repeated fury, plausible tense day +cichy_day_burstOFF_057: PASS -- public mock triggers fury, otherwise calm, recovers +cichy_day_burstOFF_058: PASS -- erupts at taunt, warm to stranger, sane recovery +cichy_day_burstOFF_059: PASS -- escalating anger to mocks, calms, recovers overnight +cichy_day_burstOFF_060: PASS -- calm day, warm to stranger, no jarring reactions \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..c9b3a68 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_051: PASS -- calm, composed; lets stranger's order pass, warm to kindness +edda_day_burstOFF_052: PASS -- settled day, gracious to help, takes charge nightly +edda_day_burstOFF_053: PASS -- ignores mock and stranger's order, calm command throughout +edda_day_burstOFF_054: PASS -- even-keeled day, warm thanks, sets watch nightly +edda_day_burstOFF_055: PASS -- composed, warmed by kindnesses, sleeps and wakes recovered +edda_day_burstOFF_056: PASS -- lets orders pass or complies, calm and settled +edda_day_burstOFF_057: PASS -- shrugs off mock, gracious, takes charge at night +edda_day_burstOFF_058: PASS -- ignores stranger's order, warm to help, sane pacing +edda_day_burstOFF_059: PASS -- settled and gracious throughout, normal sleep cycle +edda_day_burstOFF_060: PASS -- calm, warmed by repeated kindnesses, directs the watch \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..65be81f --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_051: PASS -- calm baseline, cold curtness to Wojslaw, sleeps and recovers +halgrim_day_burstOFF_052: PASS -- steady day, warm to kindness, obeys Edda, recovers +halgrim_day_burstOFF_053: PASS -- even keel, lets mockery pass, obeys Edda +halgrim_day_burstOFF_054: PASS -- calm throughout, declines Wojslaw, warmed by kindness +halgrim_day_burstOFF_055: PASS -- cold answer to public mock, mild complaint, sane +halgrim_day_burstOFF_056: PASS -- declines Wojslaw, obeys Edda, settled all day +halgrim_day_burstOFF_057: PASS -- brief tension after curt reply, settles, sleeps recovered +halgrim_day_burstOFF_058: PASS -- mostly settled, mild curtness late to Wojslaw +halgrim_day_burstOFF_059: PASS -- warm to kindness, obeys Edda, curt to Wojslaw +halgrim_day_burstOFF_060: PASS -- compliant with Edda, cold to public mock, sane diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..0d9e5f7 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_051: PASS -- restless cycles, mockery ignored, warm thanks, sleeps and wakes +lutek_day_burstOFF_052: PASS -- mild on-edge moment, otherwise calm, restless, recovers overnight +lutek_day_burstOFF_053: PASS -- novelty-seeking, insults roll off, sane pacing, sleeps well +lutek_day_burstOFF_054: PASS -- repeated mockery shrugged off, easy warmth, normal rhythm +lutek_day_burstOFF_055: PASS -- restless tinkering, calm throughout, believable day and night +lutek_day_burstOFF_056: PASS -- normal cycle, warm to kindness, unbothered by mockery +lutek_day_burstOFF_057: PASS -- steady restless rhythm, warm thanks, sane sleep timing +lutek_day_burstOFF_058: PASS -- brief on-edge, calm otherwise, insults ignored, recovers +lutek_day_burstOFF_059: PASS -- tense moment passes, novelty-seeking, mockery shrugged, sleeps +lutek_day_burstOFF_060: PASS -- mild edge late, calm man, kindness warm, recovers overnight \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..bfccf59 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_051: PASS -- sane idle-restless cycle, composed, lets mockery pass. +welf_day_burstOFF_052: PASS -- calm restless/rest rhythm, warm to kindnesses, sleeps well. +welf_day_burstOFF_053: PASS -- believable day, mild evening edge before sleep. +welf_day_burstOFF_054: PASS -- steady restlessness and ease, no jarring reactions. +welf_day_burstOFF_055: PASS -- even-keeled idle cycle, warm thanks, recovers overnight. +welf_day_burstOFF_056: PASS -- coherent pacing, composed throughout, restful night. +welf_day_burstOFF_057: PASS -- many kindnesses warmly received, calm idle rhythm. +welf_day_burstOFF_058: PASS -- believable day, brief evening tension before sleep. +welf_day_burstOFF_059: PASS -- steady restless/rest cycle, composed, recovers overnight. +welf_day_burstOFF_060: PASS -- mockery briefly bristles then settles, fitting and sane. diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..2538978 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_051: PASS -- flares at mockery, calms, sleeps and wakes recovered +wojslaw_day_burstOFF_052: PASS -- proud erupts at Marta's mockery, calm spells believable +wojslaw_day_burstOFF_053: PASS -- escalating anger then warmth, sane pacing and recovery +wojslaw_day_burstOFF_054: PASS -- mix of warmth and flares, fitting and well-paced +wojslaw_day_burstOFF_055: PASS -- erupts at insults, settles between, normal day rhythm +wojslaw_day_burstOFF_056: PASS -- warmth and fury balanced, refusals fit proud temperament +wojslaw_day_burstOFF_057: PASS -- clustered evening eruptions believable, calms before sleeping +wojslaw_day_burstOFF_058: PASS -- mild and full flares, settled stretches, recovers overnight +wojslaw_day_burstOFF_059: PASS -- mostly warm day, one mild bristle, plausible +wojslaw_day_burstOFF_060: PASS -- erupts at mockery, warm to kindness, sane arc \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_5/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..2653d2a --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_051: PASS -- Sane day; idle restlessness, warmth, evening tension, sleeps. +branic_day_burstON_052: PASS -- Believable arc; rises tense at night, wakes recovered. +branic_day_burstON_053: PASS -- Public mockery triggers snap and defiant refusals; plausible. +branic_day_burstON_054: PASS -- Evening eruption after sustained tension; otherwise composed, recovers. +branic_day_burstON_055: PASS -- Snaps at public barb; orders mostly obeyed, sleeps fine. +branic_day_burstON_056: PASS -- Calm day, lets barbs pass, normal recovery. +branic_day_burstON_057: PASS -- Late-evening fury and refusal after barb; sane pacing. +branic_day_burstON_058: PASS -- Easy day, warmth toward kindness, settled, sleeps well. +branic_day_burstON_059: PASS -- Dense orders all absorbed calmly; believable busy day. +branic_day_burstON_060: PASS -- Evening snap at order; otherwise warm and settled. diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_5/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..14cfaf3 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_051: PASS -- sane pacing, curt to guards, recovers overnight. +cichy_day_burstON_052: PASS -- believable rhythm, resentful curtness toward jailers, sleeps and recovers. +cichy_day_burstON_053: PASS -- calm day, mild bristling at mockery, fits profile. +cichy_day_burstON_054: PASS -- tense but measured, cold replies to guards, normal pacing. +cichy_day_burstON_055: PASS -- warm thanks to stranger, curt to guards, coherent. +cichy_day_burstON_056: PASS -- many guard contacts, restrained curt responses, recovers overnight. +cichy_day_burstON_057: PASS -- public mockery handled curtly, sane daily rhythm. +cichy_day_burstON_058: FLAG -- erupts in fury at a stranger's kindness, unmotivated. +cichy_day_burstON_059: PASS -- escalating mockery, snaps once, then settles believably. +cichy_day_burstON_060: PASS -- calm day, neutral to kindness, normal pacing. \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_5/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..fbe2fb6 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_051: PASS -- calm, warm, composed; lets orders pass; sane day. +edda_day_burstON_052: PASS -- settled and gracious throughout; believable pacing and rest. +edda_day_burstON_053: PASS -- ignores mockery, warm to help; coherent calm day. +edda_day_burstON_054: PASS -- composed, thankful, takes charge at night; sane. +edda_day_burstON_055: PASS -- warm and settled all day; believable rest and sleep. +edda_day_burstON_056: PASS -- lets orders pass or complies calmly; coherent day. +edda_day_burstON_057: PASS -- ignores barb, warm to kindness; composed throughout. +edda_day_burstON_058: PASS -- calm, gracious, lets order pass; sane pacing. +edda_day_burstON_059: PASS -- settled and warm; believable day and recovery. +edda_day_burstON_060: PASS -- composed, thankful, directs watch at night; coherent. diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_5/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..987976d --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_051: PASS -- calm, buttons it down; cold curtness only when pushed. +halgrim_day_burstON_052: PASS -- even keel, warm to kindness, mild curtness to Wojsław. +halgrim_day_burstON_053: PASS -- steady day, obeys Edda, lets mockery pass. +halgrim_day_burstON_054: PASS -- calm throughout, declines Wojsław, warm to small kindnesses. +halgrim_day_burstON_055: PASS -- cold contempt to Wojsław fitting; brief tension settles. +halgrim_day_burstON_056: PASS -- steady, obeys Edda, declines/ignores Wojsław plausibly. +halgrim_day_burstON_057: PASS -- mild edge after curt reply, recovers; sane pacing. +halgrim_day_burstON_058: PASS -- calm all day, cold curtness to late mockery. +halgrim_day_burstON_059: PASS -- even keel, cold reply to Wojsław, sleeps recovered. +halgrim_day_burstON_060: PASS -- obeys Edda readily, cold to public mockery, settled. \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_5/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..e97f6d9 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_051: PASS -- restless seeking, warm to kindness, shrugs mockery, sleeps well +lutek_day_burstON_052: PASS -- believable pacing, brief edge eases, sleeps and recovers +lutek_day_burstON_053: PASS -- frequent novelty-seeking, warm, mockery rolls off, sane day +lutek_day_burstON_054: PASS -- one mild curt reply, otherwise unoffended and easy +lutek_day_burstON_055: PASS -- restless cycles, warm, lets mockery pass, recovers overnight +lutek_day_burstON_056: PASS -- busy seeking, warm to kindnesses, unbothered by mockery +lutek_day_burstON_057: PASS -- mild evening edge fades, warm, sane pacing +lutek_day_burstON_058: PASS -- some evening tension but never rages, warm, sane +lutek_day_burstON_059: PASS -- restless, warm, mockery shrugged, brief edge eases +lutek_day_burstON_060: PASS -- novelty-seeking, warm, lets mockery pass, sleeps recovered diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_5/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..ae60075 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_051: PASS -- restless/eased cycle, sane sleep, mockery shrugged off +welf_day_burstON_052: PASS -- calm steady day, boredom rises and eases believably +welf_day_burstON_053: PASS -- normal idle day, settled rhythm, sleeps and recovers +welf_day_burstON_054: PASS -- composed cycling of boredom and rest throughout +welf_day_burstON_055: PASS -- even-keeled day, warm to kindnesses, sane pacing +welf_day_burstON_056: PASS -- steady boredom/ease rhythm, nothing jarring +welf_day_burstON_057: PASS -- calm day, gracious to help, believable spans +welf_day_burstON_058: PASS -- ordinary idle day, settled and recovered after sleep +welf_day_burstON_059: PASS -- composed restless/rest cycle, warm reactions, sane sleep +welf_day_burstON_060: FLAG -- snaps angrily at a kindness; contradicts even-tempered profile \ No newline at end of file diff --git a/eval/hourly_runs/slice_5/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_5/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..e05a039 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_051: PASS -- flares at mocks, defies orders, recovers, sleeps; coherent. +wojslaw_day_burstON_052: PASS -- erupts at mocks, calms between, thanks kindness; believable. +wojslaw_day_burstON_053: PASS -- escalating bristle-to-fury, warm to kindness; sane arc. +wojslaw_day_burstON_054: PASS -- mixed kindness and flares, proportionate reactions; coherent day. +wojslaw_day_burstON_055: PASS -- repeated mocks drive fury, soup mostly fine; believable. +wojslaw_day_burstON_056: PASS -- warms to help, erupts at mocks, defies orders; sane. +wojslaw_day_burstON_057: PASS -- calm morning, evening provocations spark fury; coherent. +wojslaw_day_burstON_058: PASS -- mostly settled, flares at mocks and orders; believable. +wojslaw_day_burstON_059: PASS -- mostly warm day, single flare at mock; sane pacing. +wojslaw_day_burstON_060: PASS -- erupts at mocks, warm to kindness, sleeps; coherent. diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..5fee345 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstOFF_051: PASS -- sane days, nightly reset, single justified eruption +branic_multi_burstOFF_052: PASS -- calm life, lone snap after rain, recovers +branic_multi_burstOFF_053: PASS -- pleasant calm six days, no jarring escalation +branic_multi_burstOFF_054: PASS -- mostly calm, one public eruption after rain +branic_multi_burstOFF_055: FLAG -- persistent defiance/fury snowballs, bristles at kindness, no reset +branic_multi_burstOFF_056: PASS -- calm with one night flare, resets +branic_multi_burstOFF_057: PASS -- calm with isolated flares, recovers nightly +branic_multi_burstOFF_058: FLAG -- relentless flat refusals and eruptions, no overnight reset +branic_multi_burstOFF_059: PASS -- refusals build but interspersed calm, fury isolated +branic_multi_burstOFF_060: PASS -- refusals cluster on one stressed rainy day diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..e25eab1 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstOFF_051: PASS -- sane pacing; calm baseline, mocks anger, soup ignored +cichy_multi_burstOFF_052: PASS -- believable rhythm; recovers nightly, resentment fits profile +cichy_multi_burstOFF_053: PASS -- sane days; provocations flare, kindness warms, calm baseline +cichy_multi_burstOFF_054: PASS -- coherent; tension rises and falls, sleeps recovered +cichy_multi_burstOFF_055: PASS -- plausible; mocks provoke fury, otherwise settled, fits captive +cichy_multi_burstOFF_056: PASS -- sane; strangers warm him, guards resented, nightly reset +cichy_multi_burstOFF_057: PASS -- believable pacing; flares at mocks, recovers, soup curt +cichy_multi_burstOFF_058: PASS -- coherent days; tension cycles, calm baseline holds +cichy_multi_burstOFF_059: PASS -- sane; provocations anger, kindness warms, sleeps recovered +cichy_multi_burstOFF_060: FLAG -- Day2 seven near-back-to-back furious eruptions, implausibly relentless diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..d42504f --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_051: PASS -- sane pacing, composed, lets mockery pass +edda_multi_burstOFF_052: PASS -- settled, sleeps nightly, refuses strangers' orders fittingly +edda_multi_burstOFF_053: PASS -- calm, warm to kindness, composed under barbs +edda_multi_burstOFF_054: PASS -- believable spans, recovers overnight, reserved authority +edda_multi_burstOFF_055: PASS -- settled days, gracious thanks, nothing absurd +edda_multi_burstOFF_056: PASS -- composed, mild grumble at stranger's order plausible +edda_multi_burstOFF_057: PASS -- sane rhythm, declines orders, warm to help +edda_multi_burstOFF_058: PASS -- refuses strangers, composed, sleeps and recovers +edda_multi_burstOFF_059: PASS -- settled, declines unwanted orders, takes charge +edda_multi_burstOFF_060: PASS -- calm reserved command, lets barbs pass diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..9cc7e3c --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_051: PASS -- calm baseline, sleeps/recovers, mild curt reactions only. +halgrim_multi_burstOFF_052: PASS -- steady; cold contempt only after Wojsław's repeated mockery. +halgrim_multi_burstOFF_053: PASS -- rain plus Wojsław provocations build bounded edge, recovers. +halgrim_multi_burstOFF_054: PASS -- tension rises under Wojsław, resets overnight, sane pacing. +halgrim_multi_burstOFF_055: PASS -- easy day, warm to kindness, no jarring reactions. +halgrim_multi_burstOFF_056: PASS -- escalation tracks Wojsław provocations, cold-curt not raging. +halgrim_multi_burstOFF_057: PASS -- mostly settled, mild curtness, recovers nightly. +halgrim_multi_burstOFF_058: PASS -- many kindnesses, edge builds only under Wojsław barbs. +halgrim_multi_burstOFF_059: PASS -- calm throughout, warm responses, refusals to Wojsław plausible. +halgrim_multi_burstOFF_060: PASS -- heavy Wojsław provocation, bounded cold tension, resets overnight. diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..c6346f3 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstOFF_051: PASS -- restless cycles, warm thanks, mockery shrugged, sleeps well +lutek_multi_burstOFF_052: PASS -- sane pacing, hard-to-offend, recovers nightly +lutek_multi_burstOFF_053: PASS -- novelty-seeking loop, calm, insults roll off +lutek_multi_burstOFF_054: PASS -- believable days, mockery ignored, warm to kindness +lutek_multi_burstOFF_055: PASS -- restless and good-humoured, sleeps and wakes recovered +lutek_multi_burstOFF_056: PASS -- consistent calm poet, thick-skinned, sane rhythm +lutek_multi_burstOFF_057: PASS -- easy warm cycles, no jarring reactions +lutek_multi_burstOFF_058: PASS -- restless busy days, insults shrugged, rests nightly +lutek_multi_burstOFF_059: PASS -- sane pacing, warm, mild tension only +lutek_multi_burstOFF_060: PASS -- novelty-driven, good-humoured, sleeps and recovers diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..8937422 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_051: PASS -- calm boredom cycle, sleeps and recovers, mockery shrugged off +welf_multi_burstOFF_052: PASS -- steady restlessness-and-rest rhythm, even-keeled, mockery ignored +welf_multi_burstOFF_053: PASS -- believable idle-merchant days, warm thanks, no overreaction +welf_multi_burstOFF_054: PASS -- sane pacing, composed throughout, kindnesses warmly received +welf_multi_burstOFF_055: PASS -- six days of plausible boredom cycling, recovers nightly +welf_multi_burstOFF_056: PASS -- consistent calm idleness, sleeps and wakes settled +welf_multi_burstOFF_057: PASS -- normal restless-rest rhythm, even temper holds +welf_multi_burstOFF_058: PASS -- mockery let pass, composed, believable daily spans +welf_multi_burstOFF_059: PASS -- sane rain-and-boredom days, warm and unrattled +welf_multi_burstOFF_060: PASS -- six days steady, repeated mockery shrugged, recovers each night diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..2b58223 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstOFF_051: PASS -- sane cadence, warm to kindness, flares at mockery, sleeps nightly +wojslaw_multi_burstOFF_052: PASS -- believable arc, recovers each night, anger fits proud temper +wojslaw_multi_burstOFF_053: PASS -- escalating mockery days plausible, resets overnight, fitting profile +wojslaw_multi_burstOFF_054: PASS -- normal pacing, warm-then-prickly, no absurd persistence +wojslaw_multi_burstOFF_055: PASS -- calm baseline with provoked flares, sleeps and recovers +wojslaw_multi_burstOFF_056: PASS -- heavy mockery met with fury, refusals fit pride +wojslaw_multi_burstOFF_057: PASS -- sane day/night rhythm, defiant refusals consistent with rank +wojslaw_multi_burstOFF_058: PASS -- believable buildup of temper, nightly reset, profile-consistent +wojslaw_multi_burstOFF_059: PASS -- flares at insults, warm to soup, recovers overnight +wojslaw_multi_burstOFF_060: PASS -- prickly proud noble, provoked anger only, sleeps nightly diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..b13ab90 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstON_051: PASS -- calm days, one late flare, sleeps and recovers +branic_multi_burstON_052: PASS -- mostly settled, single eruption after rain, recovers +branic_multi_burstON_053: PASS -- calm six days, mild edge, nightly reset +branic_multi_burstON_054: PASS -- gradual escalation, warmth interspersed, sleep resets +branic_multi_burstON_055: PASS -- repeated defiance but bounded, recovers each night +branic_multi_burstON_056: PASS -- calm days, single eruption, recovers normally +branic_multi_burstON_057: PASS -- builds to evening flare, sleeps and resets +branic_multi_burstON_058: FLAG -- relentless multi-day defiant refusals to respected mentor +branic_multi_burstON_059: PASS -- escalation mixed with warmth, recovers nightly +branic_multi_burstON_060: PASS -- late escalation with refusals, bounded, recovers diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..35108e9 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_051: PASS -- sane daily rhythm, cool to guards, recovers nightly +cichy_multi_burstON_052: PASS -- believable pacing, bounded flares, warm to strangers +cichy_multi_burstON_053: PASS -- coherent days, resentful to guards, anger settles +cichy_multi_burstON_054: PASS -- normal cycle, curt with captors, sleeps and recovers +cichy_multi_burstON_055: PASS -- sane spans, eruptions bounded, nightly reset intact +cichy_multi_burstON_056: PASS -- believable rhythm, warmth to strangers, cool to guards +cichy_multi_burstON_057: PASS -- coherent three days, curt responses, recovers overnight +cichy_multi_burstON_058: PASS -- sane pacing, bounded fury, settles and sleeps +cichy_multi_burstON_059: PASS -- normal days, resentful to captors, nightly recovery +cichy_multi_burstON_060: PASS -- dense mock-cluster Day2 but bounded, resets overnight diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..a64f53b --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_051: PASS -- Calm composed castellan, sane pacing, lets mockery pass. +edda_multi_burstON_052: PASS -- Sane days; declines/complains at stranger orders, fits authority. +edda_multi_burstON_053: PASS -- Steady warm command, recovers nightly, nothing jars. +edda_multi_burstON_054: PASS -- Believable spans, composed under insults, eats and rests. +edda_multi_burstON_055: PASS -- Calm authoritative days, kindness warmly received, sane sleep. +edda_multi_burstON_056: PASS -- Declines/grumbles at stranger orders only; composed otherwise. +edda_multi_burstON_057: PASS -- Steady, lets mockery pass, declines stranger order, sane. +edda_multi_burstON_058: PASS -- Refuses stranger orders, calm with own folk, recovers. +edda_multi_burstON_059: PASS -- Composed castellan, declines outside orders, believable pacing. +edda_multi_burstON_060: PASS -- Calm, lets barbs pass, runs watch, sane nights. diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..f5de1d4 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_051: PASS -- calm, mild provoked curtness, sleeps and recovers +halgrim_multi_burstON_052: PASS -- steady, escalates only under repeated Wojslaw mockery +halgrim_multi_burstON_053: PASS -- rain plus provocation drives cold contempt, recovers +halgrim_multi_burstON_054: FLAG -- cold contempt at Marta bringing soup (kindness) +halgrim_multi_burstON_055: PASS -- calm and warm throughout, well-paced +halgrim_multi_burstON_056: FLAG -- cold contempt at Marta's soup kindness +halgrim_multi_burstON_057: PASS -- provoked tension, settles by next day +halgrim_multi_burstON_058: PASS -- warm baseline, escalates under heavy Wojslaw goading +halgrim_multi_burstON_059: PASS -- calm, warm, sane pacing across days +halgrim_multi_burstON_060: FLAG -- curt/cold reply to Marta's soup kindness diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..35f75d0 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstON_051: PASS -- sane cycles, restless poet, mockery shrugged off +lutek_multi_burstON_052: PASS -- believable days; one mild curt reply, acceptable +lutek_multi_burstON_053: PASS -- coherent pacing, warm to kindness, insults pass +lutek_multi_burstON_054: FLAG -- "curtly and coldly" at mockery jars hard-to-offend profile +lutek_multi_burstON_055: PASS -- sane day/night, boredom loops, insults shrugged off +lutek_multi_burstON_056: PASS -- believable pacing, good-humoured, mockery let pass +lutek_multi_burstON_057: PASS -- coherent, warm, restless; nothing jars +lutek_multi_burstON_058: PASS -- sane cycles, novelty-seeking, insults pass +lutek_multi_burstON_059: PASS -- believable days and nights, easy-tempered throughout +lutek_multi_burstON_060: PASS -- coherent pacing, warm, restless poet; consistent diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..d99c5d4 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstON_051: PASS -- boredom cycles, sleeps and recovers, lets mockery pass +welf_multi_burstON_052: PASS -- calm restless-busy rhythm, mockery shrugged off, sane sleep +welf_multi_burstON_053: PASS -- composed, idleness-driven restlessness, recovers nightly, nothing jars +welf_multi_burstON_054: PASS -- even-keeled, rain mild tension, mockery ignored, good pacing +welf_multi_burstON_055: PASS -- steady boredom-occupy loop over six days, sleeps well +welf_multi_burstON_056: PASS -- calm rhythm, brief edge after rain, recovers fully +welf_multi_burstON_057: PASS -- composed and warm to kindness, sane daily pacing +welf_multi_burstON_058: PASS -- repeated mockery all let pass, calm throughout, recovers +welf_multi_burstON_059: PASS -- even-tempered, rain-tension mild, believable boredom cycles +welf_multi_burstON_060: PASS -- mockery shrugged off, steady restless-busy rhythm, sleeps nightly diff --git a/eval/hourly_runs/slice_5/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_5/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..1a1c396 --- /dev/null +++ b/eval/hourly_runs/slice_5/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_051: PASS -- calm baseline, erupts at mocks, recovers nightly +wojslaw_multi_burstON_052: PASS -- escalation across days, sane recovery and provocations +wojslaw_multi_burstON_053: PASS -- eruptions cluster on insults, warms when calm +wojslaw_multi_burstON_054: PASS -- dense anger day stays provocation-driven, recovers overnight +wojslaw_multi_burstON_055: PASS -- mostly calm, warm; flares tied to mocks +wojslaw_multi_burstON_056: PASS -- heavy provocation, coherent spillover, nightly reset +wojslaw_multi_burstON_057: PASS -- coherent escalation, refuses orders, recovers +wojslaw_multi_burstON_058: PASS -- warm early, anger builds on insults, recovers +wojslaw_multi_burstON_059: PASS -- flares only when already tense, sane pacing +wojslaw_multi_burstON_060: PASS -- provocation-driven eruptions, recovers nightly, coherent diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..e1210ac --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_061: PASS -- sane idle/work/rest rhythm, sleeps and recovers +branic_day_burstOFF_062: PASS -- believable day, shrugs off mockery, warm to kindness +branic_day_burstOFF_063: PASS -- restlessness and rest cycle, nothing jarring +branic_day_burstOFF_064: PASS -- brief tense moment fits thin-skinned recruit +branic_day_burstOFF_065: PASS -- repeated mockery let pass, sane pacing +branic_day_burstOFF_066: PASS -- normal rhythm, mild evening edge plausible +branic_day_burstOFF_067: PASS -- idle/work/eat/rest cycle, believable +branic_day_burstOFF_068: PASS -- dense interactions but reactions all sane +branic_day_burstOFF_069: PASS -- single bristle at mocking, otherwise calm, fits +branic_day_burstOFF_070: PASS -- believable day, warm to kindness, sleeps well diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..c3e65b6 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_061: PASS -- believable day; resentful flares at mockery, recovers, sleeps +cichy_day_burstOFF_062: PASS -- sane pacing; warm to strangers, curt to guards +cichy_day_burstOFF_063: PASS -- coherent day; fury at mockery, settles, recovers overnight +cichy_day_burstOFF_064: PASS -- believable; tense stretches, curt to guards, warm to stranger +cichy_day_burstOFF_065: PASS -- calm day, curt soup reactions, sane sleep cycle +cichy_day_burstOFF_066: PASS -- believable pacing; late flare at mockery before sleep +cichy_day_burstOFF_067: PASS -- coherent; fury at mockery, curt to guards, recovers +cichy_day_burstOFF_068: PASS -- calm day, curt to guards, sane rhythm +cichy_day_burstOFF_069: PASS -- early fury at mockery, settles, normal night +cichy_day_burstOFF_070: PASS -- believable; flare at mockery, curt soup, recovers diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..fe54788 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_061: PASS -- sane day, composed under mockery, sleeps and wakes +edda_day_burstOFF_062: PASS -- believable pacing, warm to kindness, lets stranger order pass +edda_day_burstOFF_063: PASS -- composed, one curt reply fits, normal rest cadence +edda_day_burstOFF_064: PASS -- calm day, regular meals, command at watch-setting +edda_day_burstOFF_065: PASS -- steady and at ease, kindnesses warmly received +edda_day_burstOFF_066: PASS -- composed under mockery and orders, normal sleep recovery +edda_day_burstOFF_067: PASS -- quiet sane day, lets barb pass, takes charge +edda_day_burstOFF_068: PASS -- many kindnesses, composed under barb, believable rhythm +edda_day_burstOFF_069: PASS -- calm, mockery let pass, regular meals and rest +edda_day_burstOFF_070: PASS -- sparse calm day, sane pacing, sleeps and wakes diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..3c359cb --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_061: PASS -- calm steady day, warm to kindness, sleeps well +halgrim_day_burstOFF_062: PASS -- lets mocks pass, one curt reply, even keel +halgrim_day_burstOFF_063: PASS -- mounting edge, curt then refuses Wojsław, plausible +halgrim_day_burstOFF_064: PASS -- cold curtness at Wojsław, otherwise settled and steady +halgrim_day_burstOFF_065: PASS -- placid uneventful day, calm throughout, sleeps soundly +halgrim_day_burstOFF_066: PASS -- warm and settled, lets Wojsław mock pass +halgrim_day_burstOFF_067: PASS -- calm day, refuses Wojsław's order, consistent with resentment +halgrim_day_burstOFF_068: PASS -- curt to Wojsław, obeys Edda, even tempered +halgrim_day_burstOFF_069: PASS -- settled day, obeys Edda, warm to Marta +halgrim_day_burstOFF_070: PASS -- many kindnesses, warm and settled, lets mock pass diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..c19f050 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_061: PASS -- restless cycles, mockery shrugged off, sane day/night rhythm +lutek_day_burstOFF_062: PASS -- warm to kindness, restless-busy loop, sleeps and recovers +lutek_day_burstOFF_063: PASS -- insults pass, easy good humour, believable pacing +lutek_day_burstOFF_064: PASS -- novelty-seeking cycles, warm thanks, normal sleep recovery +lutek_day_burstOFF_065: PASS -- restless and amused, kindnesses warmly received, sane rhythm +lutek_day_burstOFF_066: PASS -- mockery shrugged, busy-rest cycles, ordinary day and night +lutek_day_burstOFF_067: PASS -- insults roll off, restless tinkering, recovers after sleep +lutek_day_burstOFF_068: PASS -- late sleep but coherent, warm, restless-busy loop holds +lutek_day_burstOFF_069: PASS -- warm to help, restless cycles, late but sane sleep +lutek_day_burstOFF_070: PASS -- mockery passes, novelty-seeking, sleeps and wakes recovered diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..d9a9db7 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_061: PASS -- sane boredom/rest cycle, sleeps and recovers, mockery shrugged off +welf_day_burstOFF_062: PASS -- believable idle day, warm to kindness, lets barb pass +welf_day_burstOFF_063: PASS -- steady pacing, eats and rests, composed throughout +welf_day_burstOFF_064: PASS -- normal day, restlessness and ease alternate naturally +welf_day_burstOFF_065: PASS -- two mocks shrugged off, calm idle day, sleeps fine +welf_day_burstOFF_066: PASS -- public mockery ignored, settled mostly, believable arc +welf_day_burstOFF_067: PASS -- coherent day, warm thanks, mild evening edge, recovers +welf_day_burstOFF_068: PASS -- calm idle day, eats, rests, sleeps and recovers +welf_day_burstOFF_069: PASS -- believable pacing, neutral and warm reactions, sound sleep +welf_day_burstOFF_070: PASS -- normal boredom cycle, mild tension, recovers overnight diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..2fb5979 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_061: PASS -- sane pacing; flares at mocks, calm otherwise, sleeps +wojslaw_day_burstOFF_062: PASS -- believable escalation and recovery across the day +wojslaw_day_burstOFF_063: PASS -- anger rises and subsides plausibly; warm to kindness +wojslaw_day_burstOFF_064: PASS -- mostly warmed, erupts at public mockery; fitting +wojslaw_day_burstOFF_065: PASS -- morning flares, settles, evening eruption; coherent +wojslaw_day_burstOFF_066: PASS -- proud flares at mocks, gracious to kindness +wojslaw_day_burstOFF_067: PASS -- calm day, declines orders, late eruptions believable +wojslaw_day_burstOFF_068: PASS -- grumbles at orders, erupts at mockery; sane +wojslaw_day_burstOFF_069: PASS -- morning fury, lingering edge, recovers; plausible +wojslaw_day_burstOFF_070: PASS -- refuses orders, flares at mocks, settles; coherent diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_6/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..25c4708 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_061: PASS -- sane cycle, sleeps and wakes recovered +branic_day_burstON_062: PASS -- calm watch, lets barbs pass while worn +branic_day_burstON_063: PASS -- restless boredom cycle, late sleep but coherent +branic_day_burstON_064: PASS -- warm to kindness, believable boredom pacing +branic_day_burstON_065: PASS -- one bristle at provocation, otherwise settled +branic_day_burstON_066: PASS -- ordinary watch day, tension rises then sleeps +branic_day_burstON_067: PASS -- coherent boredom and warmth, evening edge +branic_day_burstON_068: PASS -- snaps once when worn down, plausible +branic_day_burstON_069: PASS -- flares at mockery when tense, sane otherwise +branic_day_burstON_070: PASS -- warm, busy, believable day-night rhythm diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_6/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..71c1d26 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_061: PASS -- sane pacing; resentful curtness and one flare-up fit profile +cichy_day_burstON_062: PASS -- believable day; warm to strangers, curt to guards +cichy_day_burstON_063: PASS -- calm day, cold reply to mockery, sleeps and recovers +cichy_day_burstON_064: PASS -- normal cadence; curt with guards, warm to stranger +cichy_day_burstON_065: PASS -- sane rhythm; guard kindness unsoftened, stranger thanked +cichy_day_burstON_066: PASS -- believable; late flare at mockery, recovers overnight +cichy_day_burstON_067: PASS -- steady tense day, curt guard replies, fitting +cichy_day_burstON_068: PASS -- calm sane day, curt to soup-bringer, recovers +cichy_day_burstON_069: PASS -- early fury at mockery settles, sane recovery +cichy_day_burstON_070: PASS -- normal pacing, snaps at mockery, soothes overnight diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_6/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..b85bdb1 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_061: PASS -- calm, composed; sane pacing, recovers after sleep. +edda_day_burstON_062: PASS -- warm, settled day; believable rhythm and recovery. +edda_day_burstON_063: PASS -- brief curt reply, mild edge, otherwise composed. +edda_day_burstON_064: PASS -- settled, warm responses; sane day and recovery. +edda_day_burstON_065: PASS -- composed and at ease throughout; believable pacing. +edda_day_burstON_066: PASS -- calm, lets mockery pass; sane day, recovers. +edda_day_burstON_067: PASS -- settled, warm; believable rhythm and overnight recovery. +edda_day_burstON_068: PASS -- composed, warm to kindness; sane pacing throughout. +edda_day_burstON_069: PASS -- calm, lets jabs pass; believable day, recovers. +edda_day_burstON_070: PASS -- quiet, settled day; sane pacing and recovery. diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_6/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..5414862 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_061: PASS -- calm steady day, warm to kindness, lets mockery pass +halgrim_day_burstON_062: PASS -- even keel, only mild curtness to repeated mockery +halgrim_day_burstON_063: PASS -- builds tension, cold curtness and refusal to Wojslaw, recovers +halgrim_day_burstON_064: PASS -- cold contempt to mockery only, otherwise steady, recovers overnight +halgrim_day_burstON_065: PASS -- placid day, warm thanks, mockery shrugged off +halgrim_day_burstON_066: PASS -- calm throughout, warm responses, mockery let pass +halgrim_day_burstON_067: PASS -- steady, refuses Wojslaw order at night, otherwise composed +halgrim_day_burstON_068: PASS -- mild curtness under pressure, never raised, recovers +halgrim_day_burstON_069: PASS -- entirely calm, obeys Edda, warm to Marta +halgrim_day_burstON_070: PASS -- placid warm day, mockery shrugged off, sleeps recovered diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_6/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..e160a4c --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_061: PASS -- restless cycles, shrugs off mockery, sleeps and recovers +lutek_day_burstON_062: PASS -- warm, busy, easy day; sane pacing throughout +lutek_day_burstON_063: PASS -- novelty-seeking loops, mockery ignored, normal sleep +lutek_day_burstON_064: PASS -- believable busy/rest rhythm, warm to kindnesses +lutek_day_burstON_065: PASS -- restless poet, gracious thanks, recovers overnight +lutek_day_burstON_066: PASS -- mockery passes, calm busy day, sane sleep +lutek_day_burstON_067: PASS -- insults roll off, settled mostly, normal cadence +lutek_day_burstON_068: PASS -- late sleep but coherent busy/rest day +lutek_day_burstON_069: PASS -- warm, restless, late night, believable arc +lutek_day_burstON_070: PASS -- mockery ignored, mild edge, recovers overnight diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_6/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..a99dbf7 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_061: PASS -- sane boredom cycle, sleeps and wakes, mild evening tension +welf_day_burstON_062: PASS -- lets mockery pass, calm busy/rest rhythm, recovers overnight +welf_day_burstON_063: PASS -- composed throughout, restlessness handled, mockery shrugged off +welf_day_burstON_064: PASS -- even-keeled day, believable pacing, sleeps and wakes settled +welf_day_burstON_065: PASS -- two jibes ignored, steady boredom cycle, normal sleep +welf_day_burstON_066: PASS -- public mockery shrugged, calm rhythm, recovers overnight +welf_day_burstON_067: PASS -- warm to kindness, mockery passes, sane day and rest +welf_day_burstON_068: PASS -- believable idleness cycle, mild evening edge, recovers in sleep +welf_day_burstON_069: PASS -- composed, frequent kindnesses welcomed, normal sleep and waking +welf_day_burstON_070: PASS -- steady busy/rest cadence, mild tension, recovers overnight diff --git a/eval/hourly_runs/slice_6/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_6/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..901b098 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_061: PASS -- proud flares at mocks, calm otherwise, sleeps and recovers +wojslaw_day_burstON_062: PASS -- escalating temper to mocks, settles by night, sane pacing +wojslaw_day_burstON_063: PASS -- aroused fury spills onto soup but plausible while tense +wojslaw_day_burstON_064: PASS -- mostly warmed, flares at public mock, believable proud man +wojslaw_day_burstON_065: PASS -- hot day of eruptions, tense throughout, recovers overnight +wojslaw_day_burstON_066: PASS -- calm with flares at mocks, sane rise and fall +wojslaw_day_burstON_067: PASS -- mostly settled, erupts at mocks late, recovers by morning +wojslaw_day_burstON_068: PASS -- calm, single eruption at public mock, believable noble +wojslaw_day_burstON_069: PASS -- morning fury at mocks, fades through day, recovers +wojslaw_day_burstON_070: PASS -- flares at mocks, refuses orders, settles by night diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..5f4e14e --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstOFF_061: PASS -- Sane cycles; anger escalates only under accumulated strain. +branic_multi_burstOFF_062: PASS -- Bored loops, defiance and recovery all believable. +branic_multi_burstOFF_063: PASS -- Calm days, occasional flares, nightly reset fit profile. +branic_multi_burstOFF_064: PASS -- Eager, busy, mostly compliant; pacing entirely plausible. +branic_multi_burstOFF_065: PASS -- Early flares then settling; thin-skinned recruit reads true. +branic_multi_burstOFF_066: PASS -- Repeated mockery builds to fury, otherwise calm; believable. +branic_multi_burstOFF_067: PASS -- Boredom, warmth, occasional eruptions; sane day rhythm. +branic_multi_burstOFF_068: PASS -- Compliant stretches and flares both fit, good recovery. +branic_multi_burstOFF_069: PASS -- Short run, rain, restlessness and flares all plausible. +branic_multi_burstOFF_070: PASS -- Warm, busy, even-tempered days; nothing jars profile. diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..992dd30 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstOFF_061: PASS -- calm baseline, snaps at guards, warms to strangers, sleeps well +cichy_multi_burstOFF_062: PASS -- believable rise/fall, resentment toward jailers, recovers nightly +cichy_multi_burstOFF_063: PASS -- coherent days, fury at mockery, kindness to strangers +cichy_multi_burstOFF_064: PASS -- sane pacing, guard-directed anger, restful nights +cichy_multi_burstOFF_065: PASS -- consistent captive resentment, calm spells, nightly reset +cichy_multi_burstOFF_066: PASS -- frequent eruptions at guards plausible for resentful prisoner +cichy_multi_burstOFF_067: PASS -- believable arc, anger at jailers, strangers warmly received +cichy_multi_burstOFF_068: PASS -- calm baseline with guard-triggered flares, sleeps and recovers +cichy_multi_burstOFF_069: PASS -- resentful captive, eruptions clustered around mockery, sane nights +cichy_multi_burstOFF_070: PASS -- coherent pacing, anger at guards, warmth to strangers diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..04233b1 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_061: PASS -- Calm composed days, sane sleep/recovery, measured reactions. +edda_multi_burstOFF_062: PASS -- Settled authority, believable pacing, provocations let pass. +edda_multi_burstOFF_063: PASS -- Six steady days, recovers nightly, nothing jars. +edda_multi_burstOFF_064: PASS -- Composed throughout, mild curtness fits authority figure. +edda_multi_burstOFF_065: PASS -- Sane day-cycle, declines/complains proportionately, calm. +edda_multi_burstOFF_066: PASS -- Steady command, brief curt reply plausible, recovers nightly. +edda_multi_burstOFF_067: PASS -- Six calm days, measured declines, believable rhythm. +edda_multi_burstOFF_068: PASS -- Composed, warm to kindness, provocations pass, sane sleep. +edda_multi_burstOFF_069: PASS -- Settled authority, proportionate reactions, recovers each night. +edda_multi_burstOFF_070: PASS -- Six steady days, calm command, nothing absurd. diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..3780bf6 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_061: PASS -- calm baseline, cold to Wojsław, recovers nightly +halgrim_multi_burstOFF_062: PASS -- steady keel, defiance only toward Wojsław, sleeps well +halgrim_multi_burstOFF_063: PASS -- busy provocation day, cold curtness fits, recovers +halgrim_multi_burstOFF_064: PASS -- quiet days, settled, warm to Marta +halgrim_multi_burstOFF_065: PASS -- calm with measured cold responses to mockery +halgrim_multi_burstOFF_066: PASS -- even-keeled, brief tension after evening provocations +halgrim_multi_burstOFF_067: PASS -- steady, cold contempt to Wojsław, sane pacing +halgrim_multi_burstOFF_068: PASS -- disciplined, compliant with Edda, cold to Wojsław +halgrim_multi_burstOFF_069: PASS -- mostly settled, mild grumbles, recovers nightly +halgrim_multi_burstOFF_070: PASS -- calm baseline, rain and mockery handled plausibly diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..79c4b06 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstOFF_061: PASS -- Restless cycles, warm thanks, mockery shrugged off, sleeps nightly. +lutek_multi_burstOFF_062: PASS -- Sane novelty-seeking rhythm; insults ignored; recovers each morning. +lutek_multi_burstOFF_063: PASS -- Believable pacing, good-humoured, barbs roll off, nightly sleep. +lutek_multi_burstOFF_064: PASS -- Restless busying, warm to kindness, unbothered by mockery. +lutek_multi_burstOFF_065: PASS -- Coherent days, occasional tension, mockery shrugged, sleeps recovered. +lutek_multi_burstOFF_066: PASS -- Sane stretch; curious busyness, warm thanks, no rage. +lutek_multi_burstOFF_067: PASS -- Believable cadence, insults pass, kindnesses warm, nightly rest. +lutek_multi_burstOFF_068: PASS -- Restless cycles, thick-skinned to barbs, sleeps and recovers. +lutek_multi_burstOFF_069: PASS -- Calm novelty-seeking days, warm thanks, mockery ignored. +lutek_multi_burstOFF_070: PASS -- Sane pacing, good-humoured, recovers nightly, nothing jars. diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..d9a31c8 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_061: PASS -- calm, boredom-driven, sleeps and recovers; nothing jars +welf_multi_burstOFF_062: PASS -- even-keeled restlessness cycle, sane pacing across days +welf_multi_burstOFF_063: PASS -- composed; lets mockery pass, warm to kindness +welf_multi_burstOFF_064: PASS -- believable boredom-and-rest rhythm, recovers nightly +welf_multi_burstOFF_065: PASS -- steady merchant, no overreactions, sane sleep cycle +welf_multi_burstOFF_066: PASS -- mockery shrugged off, restlessness resolves, normal pacing +welf_multi_burstOFF_067: PASS -- calm idleness cycle, sleeps at night, recovers +welf_multi_burstOFF_068: PASS -- transactional courtesy, boredom rises and falls believably +welf_multi_burstOFF_069: PASS -- even-tempered, warm thanks, sane daily rhythm +welf_multi_burstOFF_070: PASS -- composed, restless-then-settled, recovers after sleep diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..27515d2 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstOFF_061: PASS -- calm baseline, flares at insults, sleeps and recovers +wojslaw_multi_burstOFF_062: PASS -- proud refusals and fury, warm to kindness, sane nights +wojslaw_multi_burstOFF_063: PASS -- believable temper arc, recovers each morning +wojslaw_multi_burstOFF_064: PASS -- calm with warm thanks, occasional eruptions, sane pacing +wojslaw_multi_burstOFF_065: PASS -- prickly reactions to mocks, refuses orders, nightly reset +wojslaw_multi_burstOFF_066: PASS -- settled baseline, flares fit, recovers after sleep +wojslaw_multi_burstOFF_067: PASS -- short temper at insults, defiant refusals, sane days +wojslaw_multi_burstOFF_068: PASS -- flares and refusals consistent, calm recovery overnight +wojslaw_multi_burstOFF_069: PASS -- proud eruptions at mocks, warms little, sleeps soundly +wojslaw_multi_burstOFF_070: PASS -- believable temper and pride, recovers each morning diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..7b1ac6a --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstON_061: PASS -- sane cycles, escalating defiance to Halgrim, nightly recovery +branic_multi_burstON_062: PASS -- believable boredom/rest/sleep rhythm, proportionate flares +branic_multi_burstON_063: FLAG -- erupts in fury at Halgrim's kindness (Day5 07:10) +branic_multi_burstON_064: PASS -- calm compliant stretch, warm to soup, sane pacing +branic_multi_burstON_065: PASS -- restless cycles, occasional flares, recovers overnight +branic_multi_burstON_066: FLAG -- flares angrily at Halgrim's kindness (Day5 13:21) +branic_multi_burstON_067: PASS -- bored/restless rhythm, proportionate eruptions, nightly reset +branic_multi_burstON_068: PASS -- sane days, mounting temper at mockery, recovers +branic_multi_burstON_069: PASS -- believable pacing, rain endured, defiance builds plausibly +branic_multi_burstON_070: PASS -- warm to kindness, sane boredom cycles, nightly recovery diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..de90e95 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_061: PASS -- sane pacing, curt to guards, flares only on mocking +cichy_multi_burstON_062: PASS -- believable rise/fall, resentful toward guards, recovers nightly +cichy_multi_burstON_063: PASS -- coherent days, cold to captors, warm to strangers +cichy_multi_burstON_064: PASS -- steady cycles, curt guard responses, sleeps and recovers +cichy_multi_burstON_065: PASS -- sane tension arc, occasional flare, consistent grievance +cichy_multi_burstON_066: PASS -- believable pacing, cold to guards, eruption on public mockery +cichy_multi_burstON_067: PASS -- coherent multi-day life, resentful captive, recovers each night +cichy_multi_burstON_068: PASS -- sane rhythms, curt to guards, warm to stranger +cichy_multi_burstON_069: PASS -- believable arc, flares on mocking, nightly reset +cichy_multi_burstON_070: PASS -- sane pacing, cold to captors, kindness warmly received diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..3af2d5c --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_061: PASS -- Composed, sane pacing, sleeps and recovers; fitting authority. +edda_multi_burstON_062: PASS -- Calm and settled throughout; believable command demeanor. +edda_multi_burstON_063: PASS -- Steady six days, provocations let pass; consistent. +edda_multi_burstON_064: PASS -- Composed, one curt reply at most; believable. +edda_multi_burstON_065: PASS -- Calm, takes charge daily; mild complaints fit. +edda_multi_burstON_066: PASS -- Steady authority, occasional curt reply; sane arc. +edda_multi_burstON_067: PASS -- Composed across six days; provocations shrugged off. +edda_multi_burstON_068: PASS -- Calm and orderly; sleeps and recovers normally. +edda_multi_burstON_069: PASS -- Settled, lets mockery pass; consistent and believable. +edda_multi_burstON_070: PASS -- Steady six days, warm to kindness; fitting command. diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..c82a2d4 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_061: PASS -- calm baseline, cold curtness to Wojslaw, sleeps/recovers well. +halgrim_multi_burstON_062: PASS -- steady, escalates only under provocation, nightly reset holds. +halgrim_multi_burstON_063: PASS -- heavy Wojslaw barrage, cold contempt fits; brief Marta snap minor. +halgrim_multi_burstON_064: PASS -- calm, sane pacing, restrained reactions throughout. +halgrim_multi_burstON_065: PASS -- mostly settled, cold curtness only to Wojslaw mockery. +halgrim_multi_burstON_066: PASS -- calm baseline, escalation tied to provocation, recovers nightly. +halgrim_multi_burstON_067: PASS -- steady, cold contempt confined to Wojslaw, believable arc. +halgrim_multi_burstON_068: FLAG -- cold contempt to Marta serving soup jars slightly. +halgrim_multi_burstON_069: FLAG -- cold contempt to helpful stranger amid grumbling feels off. +halgrim_multi_burstON_070: FLAG -- snaps coldly at kind Marta twice, displaced spillover. diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..2464f30 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstON_061: PASS -- restless-busy cycle, warm to kindness, shrugs mockery, sleeps well. +lutek_multi_burstON_062: PASS -- sane pacing, sleeps nightly, wakes recovered, mockery shrugged. +lutek_multi_burstON_063: PASS -- believable boredom-rest rhythm, warm and thick-skinned throughout. +lutek_multi_burstON_064: PASS -- coherent days, kindness welcomed, mockery let pass, night sleep. +lutek_multi_burstON_065: PASS -- restless poet, mild curt reply once, otherwise easy and calm. +lutek_multi_burstON_066: PASS -- sane multi-day arc, warm responses, no absurd swings. +lutek_multi_burstON_067: PASS -- believable rhythm, sociable, unoffended, recovers each night. +lutek_multi_burstON_068: PASS -- coherent pacing, warm to help, shrugs barbs, sleeps nightly. +lutek_multi_burstON_069: PASS -- calm restless cycle, kindness welcomed, nothing jarring. +lutek_multi_burstON_070: PASS -- sane days and nights, warm, unbothered, well-paced. diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..0499567 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstON_061: PASS -- sane boredom cycles, sleeps and recovers, calm throughout +welf_multi_burstON_062: PASS -- believable pacing, restlessness rises and eases, no jarring rage +welf_multi_burstON_063: PASS -- lets mockery pass, even-keeled, normal sleep recovery +welf_multi_burstON_064: PASS -- rain wears mildly, no snapping, composed merchant +welf_multi_burstON_065: PASS -- shrugs off barb, steady boredom-occupation rhythm, sleeps nightly +welf_multi_burstON_066: PASS -- public mockery ignored, calm, believable daily cadence +welf_multi_burstON_067: PASS -- restless idleness, warm to kindness, sane sleep cycles +welf_multi_burstON_068: PASS -- composed, neutral to soup, boredom builds and eases plausibly +welf_multi_burstON_069: PASS -- even-keeled, normal hunger/temper spans, recovers overnight +welf_multi_burstON_070: PASS -- calm transactional sort, boredom cycles, nothing absurd diff --git a/eval/hourly_runs/slice_6/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_6/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..7b1cd02 --- /dev/null +++ b/eval/hourly_runs/slice_6/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_061: PASS -- proud temper flares, calm recovery, nightly reset believable +wojslaw_multi_burstON_062: PASS -- sane arc, eruptions at mocks, gracious mornings, sleeps well +wojslaw_multi_burstON_063: PASS -- warm to kindness, flares at insults, coherent pacing +wojslaw_multi_burstON_064: PASS -- mostly calm, appropriate refusals and warmth, believable days +wojslaw_multi_burstON_065: PASS -- hot temper consistent with proud prickly noble +wojslaw_multi_burstON_066: PASS -- gracious then souring under provocation, recovers nightly +wojslaw_multi_burstON_067: PASS -- accumulating temper, refusals, calm waking, sane span +wojslaw_multi_burstON_068: PASS -- prickly flares and refusals, coherent daily rhythm +wojslaw_multi_burstON_069: PASS -- frequent eruptions fit hot proud profile, sleeps resets +wojslaw_multi_burstON_070: PASS -- balanced flares and warmth, believable recovery and pacing diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..01b0e5a --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_071: PASS -- sane rhythm, restless cycles, calm day, sleeps and recovers +branic_day_burstOFF_072: PASS -- believable; late-evening snaps at repeated mockery when frayed +branic_day_burstOFF_073: PASS -- calm, varied kindnesses, lets jabs pass, normal pacing +branic_day_burstOFF_074: PASS -- public mockery erupts once when worn; otherwise composed +branic_day_burstOFF_075: PASS -- steady day, lets early jab pass, sleeps well +branic_day_burstOFF_076: PASS -- calm, warm to kindnesses, no jarring reactions +branic_day_burstOFF_077: PASS -- composed, lets repeated jabs pass, sane pacing +branic_day_burstOFF_078: PASS -- obeys orders, lets jabs pass, believable rhythm +branic_day_burstOFF_079: PASS -- calm dutiful day, restless cycles, recovers overnight +branic_day_burstOFF_080: PASS -- lets early mockery pass, steady, sleeps and wakes eased diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..7e0fded --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_071: PASS -- sane day; mocks anger, soup unmoved, sleeps and recovers +cichy_day_burstOFF_072: PASS -- believable pacing; curt to soup, fury at mock, rests +cichy_day_burstOFF_073: PASS -- coherent arc; resentment toward guards, calm overnight recovery +cichy_day_burstOFF_074: PASS -- sane day; eruptions at mocks, eased soup, normal sleep +cichy_day_burstOFF_075: PASS -- believable; mostly settled, evening flare, recovers by morning +cichy_day_burstOFF_076: PASS -- calm day, no mocks, soup unmoved, sane rhythm +cichy_day_burstOFF_077: PASS -- coherent; guard anger, stranger kindness warm, recovers overnight +cichy_day_burstOFF_078: PASS -- sane; early bristle, soup curt, stranger neutral, sleeps +cichy_day_burstOFF_079: PASS -- believable; stranger kindness warm, guards resented, calm recovery +cichy_day_burstOFF_080: PASS -- sane pacing; eruptions at mocks, eased soup, normal sleep diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..1b7fc15 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_071: PASS -- calm composed day, warm to kindness, sleeps and recovers +edda_day_burstOFF_072: PASS -- settled day, lets orders pass, sane pacing +edda_day_burstOFF_073: PASS -- composed, shrugs off mockery, normal rest and sleep +edda_day_burstOFF_074: PASS -- mockery passes unremarked, calm throughout, recovers overnight +edda_day_burstOFF_075: PASS -- many kindnesses, warm and settled, believable day +edda_day_burstOFF_076: PASS -- quiet ordered day, calm authority, sleeps well +edda_day_burstOFF_077: PASS -- settled, lets orders pass, normal pacing and sleep +edda_day_burstOFF_078: PASS -- composed day, warm thanks, orders shrugged off +edda_day_burstOFF_079: PASS -- obeys or shrugs orders, calm, recovers overnight +edda_day_burstOFF_080: PASS -- uneventful settled day, warm to kindness, sane sleep diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..9138837 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_071: PASS -- calm, warm to kindness, sleeps and recovers +halgrim_day_burstOFF_072: PASS -- steady easy day, sane pacing throughout +halgrim_day_burstOFF_073: PASS -- curt not raging at Wojsław, fitting control +halgrim_day_burstOFF_074: PASS -- mild edge from orders, recovers overnight +halgrim_day_burstOFF_075: PASS -- settled, warm responses, normal rest cycle +halgrim_day_burstOFF_076: PASS -- lets mockery pass, calm throughout +halgrim_day_burstOFF_077: PASS -- even keel, warm, sane day +halgrim_day_burstOFF_078: PASS -- cold curt reply to public mockery, fitting +halgrim_day_burstOFF_079: PASS -- calm, compliant, recovers at night +halgrim_day_burstOFF_080: PASS -- repeated jabs build tension, stays controlled diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..6f3e214 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_071: PASS -- restless cycles, mockery shrugged off, sleeps at night +lutek_day_burstOFF_072: PASS -- warm to kindness, restless pacing, normal night sleep +lutek_day_burstOFF_073: PASS -- believable busy cycles, easy warmth, sane night rest +lutek_day_burstOFF_074: PASS -- restless craving, kindnesses welcomed, wakes recovered at dawn +lutek_day_burstOFF_075: PASS -- mockery rolls off, restless tinkering, sane sleep cycle +lutek_day_burstOFF_076: PASS -- restless novelty-seeking, warm receipts, wakes recovered +lutek_day_burstOFF_077: PASS -- insults pass unbothered, restless cycles, normal night +lutek_day_burstOFF_078: PASS -- easy warmth, restless busying, sleeps and wakes recovered +lutek_day_burstOFF_079: PASS -- mockery shrugged, restless pacing, sane night rest +lutek_day_burstOFF_080: PASS -- restless cycles, kindnesses welcomed, wakes recovered at dawn diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..28305ae --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_071: PASS -- sane boredom cycle, calm, sleeps and recovers +welf_day_burstOFF_072: PASS -- warm to kindness, shrugs off mockery, believable +welf_day_burstOFF_073: PASS -- steady idleness cycle, calm reactions, sleeps well +welf_day_burstOFF_074: PASS -- composed, boredom rises and resets overnight +welf_day_burstOFF_075: PASS -- mild evening edge plausible, lets mockery pass +welf_day_burstOFF_076: PASS -- calm day, slight tension by night, fine +welf_day_burstOFF_077: PASS -- restless idleness, mild late edge, recovers +welf_day_burstOFF_078: PASS -- even-keeled, warm to help, sane pacing +welf_day_burstOFF_079: PASS -- steady boredom cycle, calm throughout, sleeps +welf_day_burstOFF_080: PASS -- composed, ignores mockery, believable day diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..692b0ce --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_071: PASS -- sane pacing, flares at mockery, recovers, sleeps well +wojslaw_day_burstOFF_072: PASS -- believable anger to insults, calms, recovers overnight +wojslaw_day_burstOFF_073: PASS -- proud flares at mocking, warm to kindness, sane +wojslaw_day_burstOFF_074: PASS -- short temper and defiance fit profile, recovers +wojslaw_day_burstOFF_075: PASS -- erupts at mockery, settles, sleeps and wakes calm +wojslaw_day_burstOFF_076: PASS -- consistent flaring at insults, warm thanks, sane day +wojslaw_day_burstOFF_077: PASS -- proud temper to mockery, recovers, believable arc +wojslaw_day_burstOFF_078: PASS -- mostly calm day, mild bristle, sane pacing +wojslaw_day_burstOFF_079: PASS -- erupts at public mockery, lets minor things pass +wojslaw_day_burstOFF_080: PASS -- short temper and refusal fit, recovers overnight diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_7/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..0a8899c --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_071: PASS -- sane boredom-rest cycle, warm thanks, calm day, sleeps and recovers +branic_day_burstON_072: PASS -- snaps only at late-night mockery; rest of day believable +branic_day_burstON_073: PASS -- restless loops, warmth at kindness, no unprovoked anger +branic_day_burstON_074: PASS -- erupts only at public mockery; otherwise calm and recovered +branic_day_burstON_075: PASS -- lets mockery pass, steady boredom-rest rhythm, sleeps well +branic_day_burstON_076: PASS -- calm cooperative day, warm to kindness, normal sleep +branic_day_burstON_077: PASS -- evening anger after mockery and order; believable escalation +branic_day_burstON_078: PASS -- bristles at mockery but stays composed; sane pacing +branic_day_burstON_079: PASS -- obedient, warm, restless cycles; quiet recovered night +branic_day_burstON_080: PASS -- lets jabs pass, calm cooperative day, normal sleep diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_7/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..5cd6256 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_071: PASS -- sane day; flares at mocks, curt to guards +cichy_day_burstON_072: PASS -- believable arc; eruption to mock, recovers, sleeps +cichy_day_burstON_073: PASS -- calm baseline, snaps at mocking, settles overnight +cichy_day_burstON_074: PASS -- fury at mock fits hot resentful captive +cichy_day_burstON_075: PASS -- tense-calm cycling, curt to guards, sane pacing +cichy_day_burstON_076: PASS -- quiet day, mild curtness to guard soup +cichy_day_burstON_077: PASS -- snaps at guard, warm to stranger kindness +cichy_day_burstON_078: PASS -- bristles at mock, indifferent to guard meals +cichy_day_burstON_079: PASS -- warm to strangers, edgy with guards, recovers +cichy_day_burstON_080: PASS -- two eruptions to mocking, believable, sleeps recovered diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_7/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..9e6dc1c --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_071: PASS -- composed castellan, warm to kindness, sleeps and recovers +edda_day_burstON_072: PASS -- calm day, provocations let pass, normal pacing +edda_day_burstON_073: PASS -- shrugs off mockery, settled, takes charge appropriately +edda_day_burstON_074: PASS -- barb ignored, warm to help, sane rhythm +edda_day_burstON_075: PASS -- many kindnesses welcomed, mockery let pass, composed +edda_day_burstON_076: PASS -- quiet settled day, warm thanks, proper command +edda_day_burstON_077: PASS -- order let pass, warmed by help, recovers overnight +edda_day_burstON_078: PASS -- composed, order ignored, kindnesses welcomed, sleeps well +edda_day_burstON_079: PASS -- obeys stranger calmly, warm, settled throughout +edda_day_burstON_080: PASS -- placid day of soup and thanks, takes charge diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_7/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..a6e78e7 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_071: PASS -- Calm, settled day; warm to kindness, recovers after sleep. +halgrim_day_burstON_072: PASS -- Quiet steady day, warm thanks, sane pacing throughout. +halgrim_day_burstON_073: PASS -- Lets jibes pass, answers Wojsław curtly and coldly; fitting. +halgrim_day_burstON_074: PASS -- Mild mutters under pressure, tense by night; controlled, believable. +halgrim_day_burstON_075: PASS -- Even keel, warm to kindness; uneventful, sane day. +halgrim_day_burstON_076: PASS -- Brief tension, mostly settled; warm thanks; recovers normally. +halgrim_day_burstON_077: PASS -- Lets Wojsław's order pass, warm otherwise; calm day. +halgrim_day_burstON_078: PASS -- Cold cutting contempt at public mockery; controlled, in-character. +halgrim_day_burstON_079: PASS -- Lets jibe pass, settled all day; warm thanks. +halgrim_day_burstON_080: PASS -- Repeated mockery builds cold curtness, tension; never shouts, recovers. diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_7/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..f895b1b --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_071: PASS -- restless cycles, mockery shrugged off, sleeps at night +lutek_day_burstON_072: PASS -- warm to kindness, mild evening edge, sane pacing +lutek_day_burstON_073: PASS -- believable busy-rest cycles, no jarring reactions +lutek_day_burstON_074: PASS -- sleeps, wakes recovered, restless again; consistent +lutek_day_burstON_075: PASS -- repeated mockery rolls off, warm thanks, calm +lutek_day_burstON_076: PASS -- normal day, wakes settled, restlessness fits profile +lutek_day_burstON_077: PASS -- mockery ignored, kindness warmed, healthy cadence +lutek_day_burstON_078: PASS -- warm to Marta, mild edge, recovers overnight +lutek_day_burstON_079: PASS -- mockery shrugged, evening tension only, sleeps fine +lutek_day_burstON_080: PASS -- calm, warm, restless; wakes recovered next morning diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_7/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..5f372bb --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_071: PASS -- Sane idleness/rest cycle, warm thanks, sleeps and wakes recovered. +welf_day_burstON_072: PASS -- Calm day, lets mockery pass, believable boredom pacing. +welf_day_burstON_073: PASS -- Even-keeled, grateful for kindnesses, sane sleep/wake. +welf_day_burstON_074: PASS -- Restlessness and rest cycle, warm reactions, nothing jars. +welf_day_burstON_075: PASS -- Composed, ignores mock, mild edge before night, recovers. +welf_day_burstON_076: PASS -- Believable boredom, gracious thanks, lets barb pass, sleeps. +welf_day_burstON_077: PASS -- Sane pacing, warm thanks, slight tension then sleep. +welf_day_burstON_078: PASS -- Idleness cycle, grateful, settled through the day. +welf_day_burstON_079: PASS -- Calm, warm to kindnesses, normal sleep/wake rhythm. +welf_day_burstON_080: PASS -- Composed, ignores mockery, grateful, believable day. diff --git a/eval/hourly_runs/slice_7/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_7/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..563a6a9 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_071: PASS -- quiet day, single snap on mockery, sane pacing +wojslaw_day_burstON_072: PASS -- two eruptions, wound-up bristle at kindness, coherent +wojslaw_day_burstON_073: PASS -- late double eruption, abrupt recovery before sleep acceptable +wojslaw_day_burstON_074: PASS -- sustained tension under repeated mockery, fits profile +wojslaw_day_burstON_075: PASS -- escalation arc coherent, recovers by evening +wojslaw_day_burstON_076: PASS -- bristles at stranger's kindness, proud man resents charity +wojslaw_day_burstON_077: PASS -- soup snap after two mockeries, recovers by evening +wojslaw_day_burstON_078: PASS -- calm day with many kindnesses, mild reactions only +wojslaw_day_burstON_079: PASS -- repeated soup snaps when tense, fits irritable proud man +wojslaw_day_burstON_080: PASS -- early eruptions, recovers after rest, late fury on fresh provocation diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..33e83db --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstOFF_071: PASS -- calm days, minor refusals without visible trigger +branic_multi_burstOFF_072: PASS -- steady low-key days, fatigue tense mornings fine +branic_multi_burstOFF_073: PASS -- rain and mocks build to end-of-day snap, coherent +branic_multi_burstOFF_074: FLAG -- day 4 repeated defiant refusals without new escalating triggers +branic_multi_burstOFF_075: PASS -- long calm run, kind-heavy day 4 but not absurd +branic_multi_burstOFF_076: PASS -- day 6 rain plus public mocks earn the outburst +branic_multi_burstOFF_077: PASS -- rain days handled, mild tension arc, no explosion +branic_multi_burstOFF_078: PASS -- quiet four days, compliant, good pacing throughout +branic_multi_burstOFF_079: PASS -- four stranger-kindnesses close together but not disqualifying +branic_multi_burstOFF_080: PASS -- morning mock plus rain explain cold soup non-reaction diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..dd4ebdf --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstOFF_071: PASS -- calm baseline, snaps at mockery, warmly takes stranger kindness +cichy_multi_burstOFF_072: PASS -- rain day escalates anger, two fury eruptions fit provocation load +cichy_multi_burstOFF_073: PASS -- slow tension build, fury on heavy provocation days, clean recovery +cichy_multi_burstOFF_074: PASS -- two mock eruptions Day 4 under rain wear, stranger warmth lands +cichy_multi_burstOFF_075: PASS -- three late-night eruptions dense but provocation count matches +cichy_multi_burstOFF_076: PASS -- Day 2 persistent tension with no incidents; quiet grinding plausible +cichy_multi_burstOFF_077: PASS -- back-to-back mocks trigger eruptions; Day 3 calmer, fits +cichy_multi_burstOFF_078: PASS -- heavy mock load drives multiple eruptions; settles between them +cichy_multi_burstOFF_079: PASS -- rain plus mockery chain escalates; kindness softens; calm Day 3 +cichy_multi_burstOFF_080: PASS -- one swallowed mock before later eruption; multi-day rhythm holds diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..539fbe8 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_071: PASS -- composed, well-fed, sane pacing across 3 days +edda_multi_burstOFF_072: PASS -- authority maintained, minor soup duplicate harmless +edda_multi_burstOFF_073: PASS -- kindness-dense but plausible, sleep and recovery clean +edda_multi_burstOFF_074: PASS -- escalating pushback after repeated provocations fits Edda +edda_multi_burstOFF_075: PASS -- calm, well-fed, mocks and orders absorbed without fuss +edda_multi_burstOFF_076: PASS -- three refusals in 90min dense but authority figure pushed repeatedly +edda_multi_burstOFF_077: PASS -- clean 4-day run, settled composure throughout +edda_multi_burstOFF_078: PASS -- one compliance then calm; mocks absorbed; sane +edda_multi_burstOFF_079: PASS -- muttered complaint recovers fast; soup repeats minor +edda_multi_burstOFF_080: PASS -- kindness-heavy day 1, then quieter; mocks met with composure diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..5eb4885 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_071: PASS -- rain day friction, Wojslaw resentment, sleep resets all sane +halgrim_multi_burstOFF_072: PASS -- calm days, mild curt answer to public mockery, consistent +halgrim_multi_burstOFF_073: PASS -- Wojslaw pressure escalates logically across days, sane +halgrim_multi_burstOFF_074: PASS -- sustained Wojslaw friction, Edda grumbles under pressure, sane +halgrim_multi_burstOFF_075: PASS -- mostly calm, late-night Edda grumble after stacked orders, sane +halgrim_multi_burstOFF_076: PASS -- iron self-control, mild curt responses, rain day tension sane +halgrim_multi_burstOFF_077: PASS -- heavy Wojslaw day, cold contempt earned step by step, sane +halgrim_multi_burstOFF_078: PASS -- compressed morning cluster plausible, calms and recovers sane +halgrim_multi_burstOFF_079: PASS -- very calm, friction brushed off, kindnesses received warmly +halgrim_multi_burstOFF_080: PASS -- slow-burn tension plausible, Wojslaw accumulation day 4 sane diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..d64ff85 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstOFF_071: PASS -- clean cycling, mockery rolled off, nights solid +lutek_multi_burstOFF_072: PASS -- flat soup reaction plausible after mock, days sane +lutek_multi_burstOFF_073: PASS -- tension spikes brief, resolve via rest and kindness +lutek_multi_burstOFF_074: PASS -- six days stable, mockery ignored, tension brief +lutek_multi_burstOFF_075: PASS -- boredom cycles regular, tension resolves each evening +lutek_multi_burstOFF_076: PASS -- neutral kindness response when content, pattern holds +lutek_multi_burstOFF_077: PASS -- second soup flat response plausible satiation +lutek_multi_burstOFF_078: PASS -- late nights fit restless poet, tension resolves +lutek_multi_burstOFF_079: PASS -- muted positives while tense is coherent, recovers +lutek_multi_burstOFF_080: PASS -- repeated late sleep consistent restless character diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..34d5abd --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_071: PASS -- steady boredom cycles, mocks let pass, sane sleep rhythm +welf_multi_burstOFF_072: PASS -- dense kindnesses but warmly absorbed, late sleep acceptable +welf_multi_burstOFF_073: PASS -- one neutral soup response early, otherwise sane and fitting +welf_multi_burstOFF_074: PASS -- rain watch, mocks let pass, boredom cycles sane +welf_multi_burstOFF_075: PASS -- brief restless-to-settled flip minor but not absurd +welf_multi_burstOFF_076: PASS -- one neutral kindness response, otherwise sane cycles +welf_multi_burstOFF_077: PASS -- double mocks let pass, many kindnesses warmly received +welf_multi_burstOFF_078: PASS -- two consecutive rest entries at 18:23 and 18:31 minor oddity +welf_multi_burstOFF_079: PASS -- very late night Day 3, still 6h sleep, sane overall +welf_multi_burstOFF_080: PASS -- restless and soup simultaneous harmless, sane cycles diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..cbb462f --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstOFF_071: PASS -- temper flares on mockery, settles between, sane pacing +wojslaw_multi_burstOFF_072: PASS -- heavy provocation Day 4, repeated eruptions, resets nightly +wojslaw_multi_burstOFF_073: PASS -- intense mockery load Day 1, erupts repeatedly but recovers +wojslaw_multi_burstOFF_074: PASS -- calm run driven by many kindnesses, one flare fits profile +wojslaw_multi_burstOFF_075: PASS -- low-provocation stretch, sparse eruptions, believable quiet days +wojslaw_multi_burstOFF_076: PASS -- high provocation density, multiple eruptions, pacing sane +wojslaw_multi_burstOFF_077: PASS -- two quick eruptions on back-to-back mocks, calms cleanly +wojslaw_multi_burstOFF_078: PASS -- Day 1 very heavy mockery yields many eruptions, structurally sound +wojslaw_multi_burstOFF_079: PASS -- Day 2 tense with refusals and bristling, Day 3 calm, sane arc +wojslaw_multi_burstOFF_080: PASS -- warm positive run, erupts when mockery arrives, coherent diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..6b98667 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstON_071: PASS -- displaced outburst Day 4, sleep resets hold, sane arc +branic_multi_burstON_072: PASS -- eruption Day 1, defiance Day 2, believable escalation +branic_multi_burstON_073: PASS -- rain wears Day 1, eruption Day 3 after buildup +branic_multi_burstON_074: PASS -- Day 4 persistent refusals heavy but post-eruption grudge plausible +branic_multi_burstON_075: PASS -- wound-tight Day 5, displaced soup-outburst credible +branic_multi_burstON_076: PASS -- Day 5 displaced outburst, Day 6 partial recovery with tension +branic_multi_burstON_077: PASS -- Day 5 rain plus tense, early soup-outburst plausible +branic_multi_burstON_078: PASS -- support-heavy run, no outburst, sane settled pacing +branic_multi_burstON_079: PASS -- late-night sleep Day 4 unusual but not absurd +branic_multi_burstON_080: PASS -- support-rich, no full outburst, wound-tight Day 4 night only diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..49d7d0c --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_071: PASS -- calm mornings, tension builds, single snap fits profile +cichy_multi_burstON_072: PASS -- rainy day stays tense, recovers overnight, sane +cichy_multi_burstON_073: PASS -- rain plus prior tension earns the morning outburst +cichy_multi_burstON_074: PASS -- multiple mocks, curt replies only, recovers nightly +cichy_multi_burstON_075: PASS -- outbursts spread across provocations over five days +cichy_multi_burstON_076: PASS -- quiet record, curt replies, no outbursts, sane +cichy_multi_burstON_077: PASS -- public mockery triggers eruption, sustained harassment fits +cichy_multi_burstON_078: PASS -- two outbursts under sustained harassment, sane pacing +cichy_multi_burstON_079: PASS -- late-night and rainy-morning outbursts, plausible for captive +cichy_multi_burstON_080: PASS -- two outbursts on high-harassment day, recovers each morning diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..3a69a42 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_071: PASS -- calm authority, regular rhythms, nothing jars +edda_multi_burstON_072: PASS -- varied incidents handled composedly, sane pacing +edda_multi_burstON_073: PASS -- heavy positive days, mocks absorbed, fits profile +edda_multi_burstON_074: PASS -- mild irritation builds then recovers, believable arc +edda_multi_burstON_075: PASS -- composed throughout, orders and mocks let pass cleanly +edda_multi_burstON_076: PASS -- escalating refusals fit castellan resisting stranger authority +edda_multi_burstON_077: PASS -- quiet steady days, no friction, sane +edda_multi_burstON_078: PASS -- occasional compliance and mocks absorbed, fits composed lead +edda_multi_burstON_079: PASS -- one mutter then settled, sane multi-day arc +edda_multi_burstON_080: PASS -- busy day one with kindnesses, calm thereafter, sane diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..91e7339 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_071: PASS -- tension builds across days, sleep resets hold, displaced aggression fits +halgrim_multi_burstON_072: PASS -- calm uneventful days, measured responses throughout +halgrim_multi_burstON_073: PASS -- Wojslaw provocation escalation, Edda curt only at peak tension +halgrim_multi_burstON_074: PASS -- multi-day grind, morning resets work, wound-tight label appropriate +halgrim_multi_burstON_075: PASS -- Edda contempt at day 4 peak is tight but state supports it +halgrim_multi_burstON_076: PASS -- quiet until rain day, tension arc clean and believable +halgrim_multi_burstON_077: PASS -- Marta contempt at day 3 peak is displaced aggression, fits model +halgrim_multi_burstON_078: PASS -- day 4 tense after Marta soup is minor oddity, overall sane +halgrim_multi_burstON_079: PASS -- very stoic run, both Wojslaw barbs absorbed, consistent with profile +halgrim_multi_burstON_080: PASS -- day 3 sustained tension without new input is mild, residual plausible diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..67cf081 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstON_071: PASS -- sane cycle, sleeps well, mocks shrugged off +lutek_multi_burstON_072: FLAG -- Marta's soup met with flat non-reaction, day 2 +lutek_multi_burstON_073: PASS -- frequent evening tension but never absurd, fits warm character +lutek_multi_burstON_074: PASS -- recurring morning tension clears with food, sane pattern +lutek_multi_burstON_075: PASS -- mild evening tension, mocks ignored, warmth intact +lutek_multi_burstON_076: FLAG -- stranger's kindness met with flat non-reaction, day 1 +lutek_multi_burstON_077: FLAG -- Marta's soup met with flat non-reaction, day 2 +lutek_multi_burstON_078: PASS -- late nights but recovers fine, warm responses throughout +lutek_multi_burstON_079: FLAG -- two kindnesses met with flat non-reaction while tense, day 4 +lutek_multi_burstON_080: PASS -- frequent evening tension resolves, warmth and rhythm intact diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..c778668 --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstON_071: PASS -- calm merchant, boredom cycles, sane pacing across 6 days +welf_multi_burstON_072: PASS -- heavy kindness cluster Day 1 plausible, night timing marginal but fine +welf_multi_burstON_073: PASS -- neutral soup reaction odd but composed man; mocks let pass correctly +welf_multi_burstON_074: FLAG -- Day 2 snap at mocker; profile says not hot-tempered, check burst threshold +welf_multi_burstON_075: PASS -- clean calm pattern, restlessness managed, all reactions fit profile +welf_multi_burstON_076: FLAG -- Day 1 18:59 neutral reaction to kindness; Welf usually receives help warmly +welf_multi_burstON_077: PASS -- rain tension Day 2 sane, kindness clusters Day 4 dense but believable +welf_multi_burstON_078: PASS -- regular boredom-rest cycles, consistent warm reactions, nothing jars +welf_multi_burstON_079: PASS -- 6-day arc sane, late-night Day 3 fine, rain days handled with calm +welf_multi_burstON_080: PASS -- tension mornings resolve by afternoon, mocks let pass, sane pattern diff --git a/eval/hourly_runs/slice_7/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_7/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..b463def --- /dev/null +++ b/eval/hourly_runs/slice_7/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_071: PASS -- temper builds daily, sleep resets, public insults erupt +wojslaw_multi_burstON_072: PASS -- erupts at mockings, quick evening settle after burst plausible +wojslaw_multi_burstON_073: PASS -- heavy provocation day one, volatile but sleep resets hold +wojslaw_multi_burstON_074: PASS -- days 3-5 warm due to absent provocation, not contradictory +wojslaw_multi_burstON_075: PASS -- low-provocation run, rare eruptions fit profile fine +wojslaw_multi_burstON_076: PASS -- sustained hostile environment, consistent volatility and resets +wojslaw_multi_burstON_077: PASS -- calm day one, builds to displaced aggression by day three +wojslaw_multi_burstON_078: PASS -- intense multi-day burst scenario, resets each morning intact +wojslaw_multi_burstON_079: PASS -- day three calm after rough day two, sleep reset plausible +wojslaw_multi_burstON_080: PASS -- sparse provocation yields warm stretch, erupts when mocked publicly diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..9aa89af --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_081: PASS -- restless young recruit, settles and sleeps normally +branic_day_burstOFF_082: PASS -- mocks absorbed, late barb earns a bristle, coherent arc +branic_day_burstOFF_083: PASS -- quiet day, boredom cycles, eases to sleep naturally +branic_day_burstOFF_084: PASS -- repeated mocks build to a snap, Marta's soup soothes +branic_day_burstOFF_085: PASS -- many kindnesses keep him warm; late edge fades by sleep +branic_day_burstOFF_086: FLAG -- order at 08:42 gets no reaction; recruit ignoring duty odd +branic_day_burstOFF_087: PASS -- steady day, kindness from stranger feels plausible, sleeps well +branic_day_burstOFF_088: PASS -- two soups close together odd but day arc is sane +branic_day_burstOFF_089: PASS -- orders ignored while calm but late barb stings; coherent +branic_day_burstOFF_090: PASS -- public mock shrugged off early, edge rises naturally by night diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..06063f2 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_081: PASS -- calm day, guard curtness and stranger warmth both fit +cichy_day_burstOFF_082: PASS -- quiet uneventful day, sane pacing throughout +cichy_day_burstOFF_083: PASS -- escalation to fury under mockery when already tense, believable +cichy_day_burstOFF_084: PASS -- repeated mockery drives two eruptions, proportionate responses +cichy_day_burstOFF_085: PASS -- low-incident day, mild tension arc, sane +cichy_day_burstOFF_086: PASS -- snaps at mock, lets later soup pass; anger dissipated naturally +cichy_day_burstOFF_087: PASS -- morning and afternoon eruptions under provocation, clean recovery +cichy_day_burstOFF_088: PASS -- stranger warmth then guard eruption, believable contrast +cichy_day_burstOFF_089: PASS -- morning fury despite prior kindness; guard resentment dominates +cichy_day_burstOFF_090: PASS -- unusually mild but no contradictions; low-provocation day diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..27d6ea1 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_081: PASS -- calm composed day, authority shown at night, sane pacing +edda_day_burstOFF_082: PASS -- absorbs mockery and orders without reaction, fits composed authority +edda_day_burstOFF_083: PASS -- complies with stranger order early, otherwise settled and composed +edda_day_burstOFF_084: PASS -- ignores stranger order and mock, recovers via kindnesses, sane +edda_day_burstOFF_085: PASS -- well-paced day, mocks absorbed calmly, warmth to kindnesses fits +edda_day_burstOFF_086: PASS -- ignores order and mock, steady throughout, authority at night +edda_day_burstOFF_087: PASS -- curt reply to public mock is plausible for composed authority +edda_day_burstOFF_088: PASS -- two soups in 22min odd but not a character breach, sane overall +edda_day_burstOFF_089: PASS -- many kindnesses, stays settled, authority at close, sane pacing +edda_day_burstOFF_090: PASS -- ignores mock and order without reaction, composed throughout diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..baa08e9 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_081: PASS -- slight tension late evening, resolves overnight; fits profile +halgrim_day_burstOFF_082: PASS -- calm day, tolerates Wojsław early, steady throughout +halgrim_day_burstOFF_083: PASS -- curt under Wojsław provocation, recovers; consistent +halgrim_day_burstOFF_084: PASS -- quiet uneventful day, calm and compliant; fits veteran +halgrim_day_burstOFF_085: PASS -- declines Wojsław orders; lets mock pass; fits iron control +halgrim_day_burstOFF_086: PASS -- very quiet day, steady and settled throughout +halgrim_day_burstOFF_087: FLAG -- tension persists hours after Wojsław clears; 16:02 "settled" then tense again +halgrim_day_burstOFF_088: PASS -- warm day with kindnesses; calm and compliant throughout +halgrim_day_burstOFF_089: PASS -- steady, responsive to kindness; Wojsław ignored; fits profile +halgrim_day_burstOFF_090: PASS -- two Wojsław barbs raise tension; lingers but plausible for veteran diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..0fa4d5d --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_081: PASS -- restless cycles, kindness warmly received, settles overnight +lutek_day_burstOFF_082: PASS -- sane pacing, brief late-night edge, wakes settled +lutek_day_burstOFF_083: PASS -- mock rolls off, warmth received, evening tension resolves +lutek_day_burstOFF_084: PASS -- busy generous day, calm through evening, clean sleep +lutek_day_burstOFF_085: PASS -- restless-work cycles, evening edge resolved by soup +lutek_day_burstOFF_086: PASS -- many kindnesses, evening tension eases, wakes refreshed +lutek_day_burstOFF_087: PASS -- mocks shrugged off, late tension, wakes easy +lutek_day_burstOFF_088: PASS -- active restless day, brief tension, late settling +lutek_day_burstOFF_089: FLAG -- kindness at 21:23 met with "lets it pass" not warmth +lutek_day_burstOFF_090: PASS -- mock ignored, tension mid-afternoon, resolves cleanly diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..74e5f08 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_081: PASS -- calm man, boredom cycles, mocks ignored, sleeps well +welf_day_burstOFF_082: PASS -- steady rhythm, kindnesses received warmly, no issues +welf_day_burstOFF_083: PASS -- sane pacing, two rests close together minor but plausible +welf_day_burstOFF_084: PASS -- even day, mocks brushed off, well-fed and settled +welf_day_burstOFF_085: PASS -- two mocks both ignored, restlessness cycles natural +welf_day_burstOFF_086: PASS -- well-supported day, no boredom spikes, plausible and calm +welf_day_burstOFF_087: PASS -- boredom managed throughout, kindnesses land well, coherent +welf_day_burstOFF_088: PASS -- steady day, hunger addressed, restlessness resolved normally +welf_day_burstOFF_089: PASS -- many kindnesses, mock ignored, boredom cycles intact +welf_day_burstOFF_090: PASS -- slight edge at night after late restlessness, plausible diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..78d6b78 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_081: PASS -- calm day with one late snap; fits a proud man having good day +wojslaw_day_burstOFF_082: PASS -- morning flare, builds through evening, escalating refusals fit profile +wojslaw_day_burstOFF_083: PASS -- kind morning, brief snap mid-day, proud refusals to orders at dusk +wojslaw_day_burstOFF_084: PASS -- early fury, settles, two clean outbursts spaced believably apart +wojslaw_day_burstOFF_085: PASS -- generous morning, two evening eruptions, late refusal; pacing coherent +wojslaw_day_burstOFF_086: PASS -- warm start, two public-insult eruptions, tense late night; sensible arc +wojslaw_day_burstOFF_087: PASS -- quiet day, single evening flare after order then mock; clean reset +wojslaw_day_burstOFF_088: PASS -- accumulating provocations, escalating fury by late evening; fits well +wojslaw_day_burstOFF_089: PASS -- kind morning, afternoon fury, two evening eruptions, late snap; coherent +wojslaw_day_burstOFF_090: PASS -- multiple orders absorbed, mild flares at mocks, no outsized escalation diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_8/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..ec7b74c --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_081: PASS -- restless cycling and evening tension build naturally +branic_day_burstON_082: PASS -- snap at Halgrim after accumulated mocks fits the profile +branic_day_burstON_083: PASS -- quiet day, boredom cycles, tension rising late, sane pacing +branic_day_burstON_084: PASS -- escalating mocks leading to fury outburst is believable +branic_day_burstON_085: FLAG -- snaps at Halgrim offering kindness; anger trigger without prior provocation +branic_day_burstON_086: PASS -- fury after late-night mock following long tense evening fits +branic_day_burstON_087: PASS -- positive day, Halgrim kindness accepted, wakes fresh and calm +branic_day_burstON_088: PASS -- Marta attentive all day, tension late, no outburst, plausible +branic_day_burstON_089: PASS -- snap after late mock when already tense and tired +branic_day_burstON_090: PASS -- tense evening without outburst, mock shrugged off, reasonable restraint diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_8/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..263b6be --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_081: PASS -- calm morning to tense evening, snap fits, sleeps clean +cichy_day_burstON_082: PASS -- low-stimulus quiet day, curt at soup, no flares +cichy_day_burstON_083: PASS -- two mocks, escalating responses, eventual snap believable +cichy_day_burstON_084: PASS -- cumulative wearing-down, late snap after repeated mockery +cichy_day_burstON_085: PASS -- uneventful day, tension without provocation, settles fine +cichy_day_burstON_086: PASS -- afternoon tension, single mock curt reply, settles +cichy_day_burstON_087: PASS -- early mock snaps him; resentful captive, within range +cichy_day_burstON_088: PASS -- stranger warmth, evening mock coldly rebuffed, believable arc +cichy_day_burstON_089: FLAG -- kindness then fury eruption minutes later, too fast a swing +cichy_day_burstON_090: PASS -- tension accumulates through evening, no jarring moments diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_8/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..8548068 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_081: PASS -- calm day, fed thrice, authority at end, sleeps well +edda_day_burstON_082: PASS -- absorbs two provocations without reaction, consistent composure +edda_day_burstON_083: PASS -- early order obeyed quietly, even keel throughout, sane pacing +edda_day_burstON_084: PASS -- mock and order both let pass, warmth to kindness, normal day +edda_day_burstON_085: PASS -- generous kindnesses received, mock ignored, well-fed, no jarring notes +edda_day_burstON_086: PASS -- positive day, order dismissed, remark at night let pass, settled +edda_day_burstON_087: PASS -- curt reply to mock after two soups, proportionate and credible +edda_day_burstON_088: PASS -- two soups back-to-back early, minor oddity but no behavioral jarring +edda_day_burstON_089: PASS -- heavy kindness cluster, order dismissed calmly, composed throughout +edda_day_burstON_090: PASS -- brief tension at 14:02 after mock, resolves by 16:00, believable diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_8/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..0c992aa --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_081: PASS -- Quiet day; Edda curtness late is plausible accumulation. +halgrim_day_burstON_082: PASS -- Steady, compliant day; Wojsław grumble fits the profile. +halgrim_day_burstON_083: PASS -- Marta warmth offset by Wojsław barbs; curt responses fitting. +halgrim_day_burstON_084: PASS -- Uneventful calm day; consistent with iron self-control. +halgrim_day_burstON_085: PASS -- Declining Wojsław's order fits low regard; otherwise settled. +halgrim_day_burstON_086: PASS -- Plain duty day; no friction, no issues. +halgrim_day_burstON_087: PASS -- Heavy Wojsław pressure escalates to cold contempt; sleep resets. +halgrim_day_burstON_088: PASS -- Positive-heavy day; warmth and compliance throughout, plausible. +halgrim_day_burstON_089: PASS -- Calm day with stranger kindnesses; settled throughout. +halgrim_day_burstON_090: FLAG -- Cold contempt toward Marta after Wojsław baiting; displaced snap diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_8/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..3c5327f --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_081: PASS -- restless cycle, eats twice, settles overnight, sane pacing +lutek_day_burstON_082: PASS -- slight tension late evening, sleeps and wakes calm, fits profile +lutek_day_burstON_083: PASS -- mock rolls off him, eats well, mild evening edge clears by sleep +lutek_day_burstON_084: PASS -- well-fed, active restlessness, settles smoothly, no jarring moments +lutek_day_burstON_085: PASS -- evening tension eases after soup, wakes settled, believable day +lutek_day_burstON_086: PASS -- many small kindnesses, late tension resolves, pacing sounds natural +lutek_day_burstON_087: FLAG -- wound tight at 22:04 with no food or provocation late to explain it +lutek_day_burstON_088: PASS -- mild tension spikes, resolves; stays up late but sleeps, sane overall +lutek_day_burstON_089: FLAG -- kindness at 21:23 met with no reaction; inconsistent with warm profile +lutek_day_burstON_090: PASS -- restless but regulated, two tension blips both resolve, clean sleep diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_8/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..95d5856 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_081: PASS -- restless cycles, mocks brushed off, slight evening tension fits +welf_day_burstON_082: PASS -- steady rhythm, kindnesses received warmly, mild late tension fine +welf_day_burstON_083: PASS -- two rests in quick succession at 08:00 minor oddity but harmless +welf_day_burstON_084: PASS -- calm day, mock ignored, multiple soups, no tension; sane +welf_day_burstON_085: PASS -- two mocks both let pass, fits even-keeled profile well +welf_day_burstON_086: PASS -- no hunger spike, sociable evening, settled throughout; plausible +welf_day_burstON_087: PASS -- restless bursts resolved quickly, no food noted but no distress +welf_day_burstON_088: PASS -- steady, well-fed, kindnesses warm him, sleeps on time +welf_day_burstON_089: PASS -- mock at 19:33 while resting, slight evening tension; all fits +welf_day_burstON_090: PASS -- mild late tension after idle evening work; fits profile diff --git a/eval/hourly_runs/slice_8/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_8/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..6e46f77 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_081: PASS -- calm morning, late eruption on fresh provocation, sane arc +wojslaw_day_burstON_082: PASS -- snaps early, escalating evening, multiple triggers, recovers +wojslaw_day_burstON_083: PASS -- bristles at early mock, snaps at Marta midday, calm end +wojslaw_day_burstON_084: PASS -- two eruptions on public mockery, calms between, fits profile +wojslaw_day_burstON_085: PASS -- midday and evening eruptions on clear provocations, plausible +wojslaw_day_burstON_086: PASS -- one full eruption, graduated mutter later, believable arc +wojslaw_day_burstON_087: PASS -- quiet day, minor friction handled without major explosion +wojslaw_day_burstON_088: PASS -- accumulating evening mocks, escalates to full fury, fits +wojslaw_day_burstON_089: PASS -- two public-mockery eruptions, contained between, wakes calm +wojslaw_day_burstON_090: PASS -- positive-heavy day keeps lid on, snaps only, no full fury diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..76681e0 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstOFF_081: PASS -- calm days, single snap at late-night mock, sane recovery +branic_multi_burstOFF_082: PASS -- escalating defiance under repeated mockery, explosions earned +branic_multi_burstOFF_083: PASS -- outburst after public mock, post-humiliation refusals follow +branic_multi_burstOFF_084: PASS -- warm supportive days, single snap at mock, clean rhythm +branic_multi_burstOFF_085: PASS -- simmering defiance without explosion fits burstOFF pattern +branic_multi_burstOFF_086: PASS -- Day 2 defiance cluster heavy but resentment builds credibly +branic_multi_burstOFF_087: PASS -- late-night sleep entry odd but overall pacing sound +branic_multi_burstOFF_088: PASS -- calm compliant multi-day run, no tension issues +branic_multi_burstOFF_089: PASS -- Day 4 kindness density high but not implausible +branic_multi_burstOFF_090: PASS -- morning snap then quick recovery under rain, believable diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..612d1f2 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstOFF_081: PASS -- sane 4-day arc, guard filter consistent +cichy_multi_burstOFF_082: PASS -- rain and repeated mockings drive escalation believably +cichy_multi_burstOFF_083: PASS -- heavy day-2 provocation stack yields multiple outbursts +cichy_multi_burstOFF_084: PASS -- two late outbursts fit accumulation across the day +cichy_multi_burstOFF_085: PASS -- repeated guard pressure sustains elevated anger across days +cichy_multi_burstOFF_086: PASS -- five day-1 outbursts dense but provocation count matches +cichy_multi_burstOFF_087: PASS -- double soup two minutes apart is quirky, not a character flaw +cichy_multi_burstOFF_088: PASS -- rain plus sustained mockings yield credible outburst cluster +cichy_multi_burstOFF_089: PASS -- stranger warmth then guard eruption fits dual-filter profile +cichy_multi_burstOFF_090: PASS -- two day-4 outbursts spaced plausibly under rain stressor diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..31cdc3b --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_081: PASS -- calm authority, sane routine, warmed by kindness +edda_multi_burstOFF_082: PASS -- one compliance odd but not impossible for a castellan +edda_multi_burstOFF_083: PASS -- uneventful days, steady composure throughout +edda_multi_burstOFF_084: PASS -- settled, lets barbs pass, consistent profile +edda_multi_burstOFF_085: PASS -- early compliance low but not repeated; overall sane +edda_multi_burstOFF_086: PASS -- social days, warm responses, no jarring moments +edda_multi_burstOFF_087: PASS -- friction day handled with irritation but not rage +edda_multi_burstOFF_088: PASS -- abundant kindness, composed throughout, normal pacing +edda_multi_burstOFF_089: PASS -- warm busy days, authority asserted appropriately +edda_multi_burstOFF_090: PASS -- firm refusal to stranger fits high-standing castellan diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..d477c52 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_081: PASS -- calm run, appropriate mild friction with Wojslaw +halgrim_multi_burstOFF_082: PASS -- tension builds across days under Wojslaw pressure, resets at sleep +halgrim_multi_burstOFF_083: PASS -- quiet peaceful stretch, no provocations, stays settled +halgrim_multi_burstOFF_084: PASS -- rain plus Wojslaw friction escalates believably, resets overnight +halgrim_multi_burstOFF_085: PASS -- six-day Wojslaw grind, progressive tension, Edda spillover on Day 4 earned +halgrim_multi_burstOFF_086: PASS -- rain day absorbed, good Marta support, clean arc +halgrim_multi_burstOFF_087: PASS -- Wojslaw jabs accumulate across days, tension carries, Edda respected +halgrim_multi_burstOFF_088: FLAG -- Day 5: flat non-response to Marta's soup, breaks warm pattern +halgrim_multi_burstOFF_089: PASS -- graduated Wojslaw responses, tension rises Day 3, believable +halgrim_multi_burstOFF_090: PASS -- heavy Day 1 Wojslaw barrage, spillover to Edda on Day 3 earned diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..ec9ee79 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstOFF_081: PASS -- restless cycle, mocks roll off, sane pacing throughout +lutek_multi_burstOFF_082: PASS -- late sleep day 1 fine, brief tension plausible fatigue +lutek_multi_burstOFF_083: PASS -- sated on kindness believable after dense positive run +lutek_multi_burstOFF_084: PASS -- evening tension after long work stretch, resolves cleanly +lutek_multi_burstOFF_085: PASS -- two flat food responses consistent with satiation state +lutek_multi_burstOFF_086: PASS -- recurring evening tension resolves with rest, mocks ignored +lutek_multi_burstOFF_087: PASS -- early-morning tension pattern consistent across days, no contradictions +lutek_multi_burstOFF_088: PASS -- recurring early tension, sated responses plausible, overall sane +lutek_multi_burstOFF_089: PASS -- tense ignored soup while preoccupied minor, recovers after rest +lutek_multi_burstOFF_090: PASS -- morning and evening tension states with calm midday, believable diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..4c6468f --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_081: PASS -- steady boredom cycles, mocks ignored, sane pacing throughout +welf_multi_burstOFF_082: PASS -- composed under mocks, morning tension from early work plausible +welf_multi_burstOFF_083: PASS -- brief restless flicker at 22:36-42 minor, overall sane +welf_multi_burstOFF_084: PASS -- tense after kindness at 06:14 slightly odd but boredom dominates +welf_multi_burstOFF_085: PASS -- recurring morning tension pattern consistent across days +welf_multi_burstOFF_086: PASS -- stable six-day rhythm, recovers each morning, sane +welf_multi_burstOFF_087: PASS -- soup ignored on rain watch; all else fits composed merchant +welf_multi_burstOFF_088: PASS -- midnight sleep transition awkward but not a model failure +welf_multi_burstOFF_089: PASS -- four kindnesses close together on Day 3 dense but not absurd +welf_multi_burstOFF_090: PASS -- clean three-day record, boredom and warmth cycling normally diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..bddfe7f --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstOFF_081: PASS -- calm start escalates to fury-day, sleeps reset cleanly +wojslaw_multi_burstOFF_082: PASS -- early public insult eruption, warm recovery days follow +wojslaw_multi_burstOFF_083: PASS -- double mock triggers double eruption, tension drains by night +wojslaw_multi_burstOFF_084: PASS -- dense insult clusters on days 3-4 yield dense eruptions, sane +wojslaw_multi_burstOFF_085: PASS -- good day 3 warmth; day 4 lets mocks pass, slightly mild but plausible +wojslaw_multi_burstOFF_086: PASS -- day 2 tension builds without clear trigger, acceptable background wear +wojslaw_multi_burstOFF_087: PASS -- heavy-provocation days spike anger, quiet days stay calm +wojslaw_multi_burstOFF_088: PASS -- bristles at kindness under tension fits a proud man in foul mood +wojslaw_multi_burstOFF_089: PASS -- warm day 1, building tension days 2-3, eruptions well-provoked +wojslaw_multi_burstOFF_090: PASS -- unusually peaceful four days with no provocations, consistent diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..7e60fee --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstON_081: PASS -- clean arc, single evening burst, recovers well +branic_multi_burstON_082: PASS -- post-burst refusals Day 2 plausible, Day 5 burst earned +branic_multi_burstON_083: PASS -- burst on kindness reads as displaced aggression, borderline +branic_multi_burstON_084: PASS -- warm Day 2 recovery after Day 1 outburst convincing +branic_multi_burstON_085: FLAG -- erupts twice from plain orders not provocations, same evening +branic_multi_burstON_086: PASS -- resentment refusals and rain tension arc coherent +branic_multi_burstON_087: PASS -- soup-triggered fury is displaced aggression, burstON plausible +branic_multi_burstON_088: PASS -- burst after two tense days, good recovery arc Days 4-6 +branic_multi_burstON_089: PASS -- tension-recovery cycle across 6 days reads naturally +branic_multi_burstON_090: PASS -- morning snap at mock after tense start, clean 3-day arc diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..cbbeae4 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_081: PASS -- sane pacing, curtness to guards, nights normal +cichy_multi_burstON_082: PASS -- rain and mockings escalate appropriately, stranger warmth fine +cichy_multi_burstON_083: PASS -- multiple outbursts fit burstON, post-burst recovery plausible +cichy_multi_burstON_084: PASS -- rain days tense, stranger soothes, pacing believable +cichy_multi_burstON_085: PASS -- high outburst volume fits burstON, each guard-triggered +cichy_multi_burstON_086: PASS -- late-night eruption before sleep is odd but not absurd +cichy_multi_burstON_087: PASS -- duplicate soup at 08:16/08:18 is noise, not a break +cichy_multi_burstON_088: PASS -- 40-min post-outburst recovery is fast but plausible +cichy_multi_burstON_089: PASS -- three stranger kindnesses in one evening is heavy but acceptable +cichy_multi_burstON_090: PASS -- steady escalation under rain and mocking, sane throughout diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..38a9700 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_081: PASS -- composed castellan, sane pacing, appropriate authority throughout +edda_multi_burstON_082: PASS -- steady days, one stranger compliance in good mood, acceptable +edda_multi_burstON_083: PASS -- calm and well-supported, duplicate soup events minor artefact +edda_multi_burstON_084: PASS -- two mocks absorbed without reaction, fits composed profile +edda_multi_burstON_085: PASS -- one early compliance after warm morning, borderline but within range +edda_multi_burstON_086: PASS -- long quiet run, authority exercised regularly, nothing odd +edda_multi_burstON_087: PASS -- graduated responses to repeated stranger orders, believable +edda_multi_burstON_088: PASS -- warm and well-supported throughout, pacing sane +edda_multi_burstON_089: PASS -- dense kindnesses on day 5 but no structural issue +edda_multi_burstON_090: PASS -- firm refusal of stranger order fits self-possessed castellan diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..afd2742 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_081: PASS -- quiet days, Wojsław friction handled, sane recovery +halgrim_multi_burstON_082: PASS -- earned escalation under Wojsław pressure across 4 days +halgrim_multi_burstON_083: PASS -- very calm run, minimal provocation, fits profile well +halgrim_multi_burstON_084: PASS -- rain plus Wojsław mocking builds tension credibly +halgrim_multi_burstON_085: PASS -- long simmering state, Marta blowback reads as displaced irritability +halgrim_multi_burstON_086: PASS -- rain day absorbed well, good recovery arc across 5 days +halgrim_multi_burstON_087: PASS -- gradual Wojsław escalation, tension spill to Edda plausible +halgrim_multi_burstON_088: PASS -- multi-day strain, Marta contempt is displaced but sustained state warrants +halgrim_multi_burstON_089: PASS -- quiet build over 3 days, Wojsław tension earned gradually +halgrim_multi_burstON_090: PASS -- Wojsław barrage Day1 earns sustained tension, Edda spill plausible diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..58eb972 --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstON_081: PASS -- restless cycle steady, mockery rolls off, pacing sane +lutek_multi_burstON_082: PASS -- late night Day 1 plausible, evening tension resolves cleanly +lutek_multi_burstON_083: PASS -- stranger kindness ignored once while tense, minor, acceptable +lutek_multi_burstON_084: PASS -- morning tension without trigger, recovers well, sane arc +lutek_multi_burstON_085: PASS -- two muted soup responses when sated, minor, within range +lutek_multi_burstON_086: PASS -- curt reply to mock under sustained tension, burst-on expected +lutek_multi_burstON_087: PASS -- repeated morning tension patterns, resolves each day cleanly +lutek_multi_burstON_088: PASS -- multi-day morning tension on waking, food/kindness recovers it +lutek_multi_burstON_089: PASS -- curt reply under tension fits thick-skinned man at his limit +lutek_multi_burstON_090: PASS -- smooth arc, tension appears and resolves, nothing jars diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..000520d --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstON_welf.txt @@ -0,0 +1,12 @@ +All records reviewed. Welf is calm, boredom-driven, mostly lets mocks pass. One snap in 084 (Day 5) — single, brief, recovers; acceptable for a composed man under accumulated rain/tension, not raging all day. + +welf_multi_burstON_081: PASS -- Sane idle cycle; mocks shrugged off; sleeps and recovers. +welf_multi_burstON_082: PASS -- Brief edge after mock, mostly composed; believable boredom pacing. +welf_multi_burstON_083: PASS -- Even-keeled, restless-then-occupied loop; mocks pass without flaring. +welf_multi_burstON_084: PASS -- One mild snap after rain/tension, recovers; otherwise calm. +welf_multi_burstON_085: PASS -- Calm, boredom-driven, kindnesses warm him; nothing jars. +welf_multi_burstON_086: PASS -- Steady six-day rhythm; tension rises and resets nightly. +welf_multi_burstON_087: PASS -- Composed throughout; restless cycles and sleep recovery believable. +welf_multi_burstON_088: PASS -- Sane pacing across nights; mocks ignored, sleeps and wakes settled. +welf_multi_burstON_089: PASS -- Even-tempered; rain tension mild, mocks shrugged; coherent days. +welf_multi_burstON_090: PASS -- Calm idle loops, nightly reset; nothing contradicts the profile. \ No newline at end of file diff --git a/eval/hourly_runs/slice_8/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_8/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..5aba34c --- /dev/null +++ b/eval/hourly_runs/slice_8/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_081: PASS -- anger accumulates across day 5, sleep resets hold +wojslaw_multi_burstON_082: PASS -- early eruption fades, next days rebuild plausibly +wojslaw_multi_burstON_083: PASS -- double mock triggers burst, displaced aggression onto soup +wojslaw_multi_burstON_084: PASS -- sustained burst mode across 6 days, mornings reset +wojslaw_multi_burstON_085: PASS -- bad day 2 then calm day 3 contrast is believable +wojslaw_multi_burstON_086: PASS -- positive day 3 earned after two tense prior days +wojslaw_multi_burstON_087: PASS -- long wear-down pattern, consistent morning resets +wojslaw_multi_burstON_088: PASS -- calm day 1, escalation day 3, displaced aggression present +wojslaw_multi_burstON_089: PASS -- 3-day arc from serene to irritable, clean progression +wojslaw_multi_burstON_090: PASS -- three calm days after one burst; no provocation, plausible diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstOFF_branic.txt b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_branic.txt new file mode 100644 index 0000000..a807fd4 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstOFF_091: PASS -- Eruption only on public mock; calm pacing, sleeps, recovers. +branic_day_burstOFF_092: PASS -- Sane day, restless-bored cadence, mild edge, recovers overnight. +branic_day_burstOFF_093: PASS -- Calm cooperative day, believable boredom and rest cycle. +branic_day_burstOFF_094: PASS -- Restless young recruit, mild tension, sleeps and wakes settled. +branic_day_burstOFF_095: PASS -- Public mock erupts then refusal; plausible thin-skinned reaction. +branic_day_burstOFF_096: PASS -- Busy cooperative day, mild edge, recovers overnight. +branic_day_burstOFF_097: PASS -- Snaps then declines order after needling; coherent escalation. +branic_day_burstOFF_098: PASS -- Calm obedient day, normal boredom and rest rhythm. +branic_day_burstOFF_099: PASS -- Compliant day, mild evening edge, sleeps and recovers. +branic_day_burstOFF_100: FLAG -- Sleeps at 00:34, very late; two refusals after mild mockery. \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstOFF_cichy.txt b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_cichy.txt new file mode 100644 index 0000000..776d315 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstOFF_091: PASS -- calm day, curt at guard soup, sleeps and recovers. +cichy_day_burstOFF_092: PASS -- sane pacing, mild edge, restful night, believable throughout. +cichy_day_burstOFF_093: PASS -- flares at guard mockery, calms between, consistent resentment. +cichy_day_burstOFF_094: PASS -- erupts at public mock, otherwise settled, believable arc. +cichy_day_burstOFF_095: PASS -- quiet day, soup ignored, normal rest cycle. +cichy_day_burstOFF_096: PASS -- tense stretches, curt at guards, sane recovery overnight. +cichy_day_burstOFF_097: PASS -- one fury at mockery, calm otherwise, recovers by night. +cichy_day_burstOFF_098: PASS -- warm to strangers, furious at guards; contrast fits profile. +cichy_day_burstOFF_099: PASS -- mostly calm, warm to stranger, neutral to guards. +cichy_day_burstOFF_100: PASS -- repeated guard mockery sparks fury, calms appropriately, sleeps. \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstOFF_edda.txt b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_edda.txt new file mode 100644 index 0000000..03b05e2 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstOFF_091: PASS -- calm composed castellan, lets stranger's order pass, sleeps well +edda_day_burstOFF_092: PASS -- composed throughout, mild complaint at stranger's order, recovers +edda_day_burstOFF_093: PASS -- settled, warm to kindnesses, takes charge, sleeps recovered +edda_day_burstOFF_094: PASS -- lets mockery pass calmly, composed authority, sane pacing +edda_day_burstOFF_095: PASS -- many kindnesses, mockery shrugged off, commands watch, sleeps +edda_day_burstOFF_096: PASS -- warm and settled, directs others, believable day +edda_day_burstOFF_097: PASS -- quiet composed day, warm to help, takes charge +edda_day_burstOFF_098: PASS -- composed, mockery ignored, warm to soup, sleeps well +edda_day_burstOFF_099: PASS -- mockery shrugged off, warm to kindness, commands, recovers +edda_day_burstOFF_100: PASS -- settled, lets mockery pass, directs watch, sane pacing \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstOFF_halgrim.txt b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_halgrim.txt new file mode 100644 index 0000000..b10a135 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstOFF_091: PASS -- calm steady day, obeys Edda, warmed by kindnesses, sleeps. +halgrim_day_burstOFF_092: PASS -- shrugs off mockery, stays at ease, recovers overnight. +halgrim_day_burstOFF_093: PASS -- declines Wojsław, calm throughout, believable pacing. +halgrim_day_burstOFF_094: PASS -- obeys Edda, refuses Wojsław, even keel all day. +halgrim_day_burstOFF_095: PASS -- cold curt replies to public mockery, fits iron control. +halgrim_day_burstOFF_096: PASS -- curt to Wojsław, obeys Edda, warm to kindness. +halgrim_day_burstOFF_097: PASS -- steady dutiful day, recovers overnight, nothing jars. +halgrim_day_burstOFF_098: PASS -- mockery builds mild tension, eases by evening, plausible. +halgrim_day_burstOFF_099: PASS -- quiet uneventful day, calm throughout, wakes recovered. +halgrim_day_burstOFF_100: PASS -- mild curtness to Wojsław, obeys Edda, calm. \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstOFF_lutek.txt b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_lutek.txt new file mode 100644 index 0000000..580a143 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstOFF_091: PASS -- restless cycles, shrugs off mockery, sleeps and wakes recovered +lutek_day_burstOFF_092: PASS -- believable boredom rhythm, warm to kindness, mild tension only +lutek_day_burstOFF_093: PASS -- sane pacing, mockery rolls off, recovers overnight +lutek_day_burstOFF_094: PASS -- restless-rest cycles, thick-skinned to insult, normal sleep +lutek_day_burstOFF_095: PASS -- warm responses, frequent restlessness, calm throughout +lutek_day_burstOFF_096: PASS -- coherent day, warm to Marta, settled overall +lutek_day_burstOFF_097: PASS -- mockery shrugged off, restless rhythm, recovers overnight +lutek_day_burstOFF_098: PASS -- believable boredom cycles, warm thanks, mild evening tension +lutek_day_burstOFF_099: PASS -- calm easy day, frequent novelty-seeking, normal sleep +lutek_day_burstOFF_100: PASS -- sane pacing, warm to kindness, recovers with light \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstOFF_welf.txt b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_welf.txt new file mode 100644 index 0000000..f3ca4ea --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstOFF_091: PASS -- sane cycle of restlessness, occupation, rest, warm thanks +welf_day_burstOFF_092: PASS -- calm day, mild reactions, sleeps and wakes recovered +welf_day_burstOFF_093: PASS -- lets mockery pass calmly, fitting composed merchant +welf_day_burstOFF_094: PASS -- brief tension then calm, mockery shrugged off, believable +welf_day_burstOFF_095: PASS -- even-keeled day, restlessness and rest in good rhythm +welf_day_burstOFF_096: PASS -- normal cycle of boredom, occupation, warm responses +welf_day_burstOFF_097: PASS -- many kindnesses warmly received, sane pacing throughout +welf_day_burstOFF_098: PASS -- mild tension, mockery ignored, late sleep still reasonable +welf_day_burstOFF_099: PASS -- steady restlessness-and-rest rhythm, recovers overnight +welf_day_burstOFF_100: PASS -- ordinary even-keeled day, mild reactions, sleeps well \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_wojslaw.txt new file mode 100644 index 0000000..bbe0e25 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstOFF_091: PASS -- flares at mocks, calms between, normal pacing +wojslaw_day_burstOFF_092: PASS -- heavy provocation day, eruptions and defiance fit pride +wojslaw_day_burstOFF_093: PASS -- four eruptions all mock-triggered, sustained tension plausible +wojslaw_day_burstOFF_094: PASS -- muted day, late snap at mock, within variable-temper range +wojslaw_day_burstOFF_095: PASS -- warmth then eruption both provocation-linked, fast swings passable +wojslaw_day_burstOFF_096: PASS -- quieter day, mocks let pass, no profile violation +wojslaw_day_burstOFF_097: FLAG -- erupts at soup delivery, no provocation in the act +wojslaw_day_burstOFF_098: PASS -- inconsistent responses but within variable-temper range +wojslaw_day_burstOFF_099: PASS -- genuinely kind day, sustained warmth, no insults until late +wojslaw_day_burstOFF_100: PASS -- erupts at mock, calms quickly, responses fit stimuli diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstON_branic.txt b/eval/hourly_runs/slice_9/results_prev/day_burstON_branic.txt new file mode 100644 index 0000000..7bec2f3 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_day_burstON_091: PASS -- restless/eased cycle, sleeps and wakes recovered, sane pacing. +branic_day_burstON_092: PASS -- tension builds over idle day, sleep resets, believable. +branic_day_burstON_093: PASS -- calm cooperative day, warm to kindnesses, sleeps well. +branic_day_burstON_094: PASS -- ordinary restless day, mild evening tension, recovers overnight. +branic_day_burstON_095: PASS -- evening eruptions/refusal after mocking, plausible for thin-skinned recruit. +branic_day_burstON_096: PASS -- late-day fury after long idle wear, sleep resets. +branic_day_burstON_097: PASS -- snaps and declines after repeated mocking, believable buildup. +branic_day_burstON_098: PASS -- compliant calm day, faint evening edge, recovers overnight. +branic_day_burstON_099: PASS -- obedient day, mild night tension, wakes settled. +branic_day_burstON_100: PASS -- flares and refuses after mocking, sane temper arc. \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstON_cichy.txt b/eval/hourly_runs/slice_9/results_prev/day_burstON_cichy.txt new file mode 100644 index 0000000..d45a656 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_day_burstON_091: PASS -- calm prisoner, curt with guards, sleeps and recovers +cichy_day_burstON_092: PASS -- sane pacing, mild curtness toward guards, normal rest +cichy_day_burstON_093: PASS -- erupts at mocking, settles after; believable resentful captive +cichy_day_burstON_094: PASS -- fury at public mockery, curt with guards, recovers +cichy_day_burstON_095: PASS -- calm day, curt with soup-bringing guard, sleeps well +cichy_day_burstON_096: PASS -- mild tension, curt with guards, normal rhythm +cichy_day_burstON_097: PASS -- erupts at mockery, otherwise curt and settled +cichy_day_burstON_098: PASS -- warm to strangers, curt with guards; contrast fits profile +cichy_day_burstON_099: PASS -- calm, curt with guards, warm to stranger, sleeps +cichy_day_burstON_100: PASS -- snaps at mockery, curt with jailers, recovers overnight \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstON_edda.txt b/eval/hourly_runs/slice_9/results_prev/day_burstON_edda.txt new file mode 100644 index 0000000..ff15484 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_day_burstON_091: PASS -- composed castellan, warm to kindness, lets order pass calmly +edda_day_burstON_092: PASS -- mild mutter at stranger's order, otherwise settled; believable +edda_day_burstON_093: PASS -- calm day, warm to Marta, sane sleep/wake cadence +edda_day_burstON_094: PASS -- lets mockery pass, composed; fits reserved authority figure +edda_day_burstON_095: PASS -- brief tension at midday, recovers; warm and settled throughout +edda_day_burstON_096: PASS -- steady calm day, warm to staff, normal pacing +edda_day_burstON_097: PASS -- quiet uneventful day, composed, recovers overnight +edda_day_burstON_098: PASS -- many kindnesses, lets mockery pass, settled and composed +edda_day_burstON_099: PASS -- brief tension after mockery, recovers; calm and warm +edda_day_burstON_100: PASS -- lets mockery pass, warm to staff, sane cadence \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstON_halgrim.txt b/eval/hourly_runs/slice_9/results_prev/day_burstON_halgrim.txt new file mode 100644 index 0000000..6e5398e --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_day_burstON_091: PASS -- calm, obeys, warmed by kindness; sane day. +halgrim_day_burstON_092: PASS -- ignores Wojsław's barb, easy day, sleeps well. +halgrim_day_burstON_093: PASS -- declines arrogant Wojsław, warm to Marta; consistent. +halgrim_day_burstON_094: PASS -- obeys Edda, refuses Wojsław; calm throughout. +halgrim_day_burstON_095: PASS -- cold contempt to repeated public mockery; fitting. +halgrim_day_burstON_096: PASS -- curt cold reply to Wojsław, otherwise steady. +halgrim_day_burstON_097: PASS -- uneventful, obedient, warmly fed; sane. +halgrim_day_burstON_098: PASS -- curt to mockery, tense but controlled; recovers overnight. +halgrim_day_burstON_099: PASS -- quiet uneventful day, settled throughout. +halgrim_day_burstON_100: PASS -- obeys Edda, mild curtness to Wojsław; steady. \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstON_lutek.txt b/eval/hourly_runs/slice_9/results_prev/day_burstON_lutek.txt new file mode 100644 index 0000000..1b0c069 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_day_burstON_091: PASS -- Restless cycles, soup, mockery shrugged off, sleeps and wakes recovered. +lutek_day_burstON_092: PASS -- Warm responses, frequent novelty-seeking, calm overnight reset; fully in character. +lutek_day_burstON_093: PASS -- Mockery rolls off, easy boredom-driven activity, normal sleep cycle. +lutek_day_burstON_094: PASS -- Two slights shrugged off, restless tinkering, believable pacing throughout. +lutek_day_burstON_095: PASS -- Many kindnesses welcomed warmly, restless cadence, sane sleep recovery. +lutek_day_burstON_096: PASS -- Warm to Marta, boredom cycles, calm day, normal night. +lutek_day_burstON_097: PASS -- Mockery ignored, soup welcomed, restless activity, sleeps and recovers. +lutek_day_burstON_098: PASS -- Easy good humor, novelty-seeking, believable hunger/rest rhythm overnight. +lutek_day_burstON_099: PASS -- Warm thanks, restless tinkering, settled day, normal sleep reset. +lutek_day_burstON_100: PASS -- Kindnesses received warmly, boredom-driven activity, sane daily rhythm. \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstON_welf.txt b/eval/hourly_runs/slice_9/results_prev/day_burstON_welf.txt new file mode 100644 index 0000000..e4f36c0 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_day_burstON_091: PASS -- Calm, restless-when-idle, warms to kindness, sleeps and recovers. +welf_day_burstON_092: PASS -- Even-keeled, boredom cycles, normal pacing, restful night. +welf_day_burstON_093: PASS -- Shrugs off mocks, warms to kindness, sane day. +welf_day_burstON_094: PASS -- Mild edge once but lets mocks pass; believable. +welf_day_burstON_095: PASS -- Composed throughout, ignores mockery, normal rhythm. +welf_day_burstON_096: PASS -- Boredom cycles, gracious to help, sleeps and recovers. +welf_day_burstON_097: PASS -- Calm, well-fed, restless intervals, normal sleep. +welf_day_burstON_098: PASS -- Brief tension, late sleep, otherwise composed and sane. +welf_day_burstON_099: PASS -- Even-keeled boredom cycles, warm thanks, normal night. +welf_day_burstON_100: PASS -- Steady, courteous, idleness-driven restlessness, recovers overnight. \ No newline at end of file diff --git a/eval/hourly_runs/slice_9/results_prev/day_burstON_wojslaw.txt b/eval/hourly_runs/slice_9/results_prev/day_burstON_wojslaw.txt new file mode 100644 index 0000000..131efad --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/day_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_day_burstON_091: PASS -- displaced anger at kindness after two provocations, plausible +wojslaw_day_burstON_092: PASS -- erupts at mockery and defies orders; fury at helper after rough day ok +wojslaw_day_burstON_093: PASS -- twice-mocked then snaps at soup; evening reactions diminish sensibly +wojslaw_day_burstON_094: PASS -- calm day, lets early barb pass, erupts at afternoon insult +wojslaw_day_burstON_095: PASS -- sustained anger displacing onto kindnesses; wakes reset +wojslaw_day_burstON_096: FLAG -- fury at 20:03 then visibly warmed and grateful twice within 30 min +wojslaw_day_burstON_097: PASS -- overflow fury at soup after barb; sustained tension fits profile +wojslaw_day_burstON_098: PASS -- quiet day, muted reactions; bristles only late; sane pacing +wojslaw_day_burstON_099: FLAG -- proud prickly man warmly grateful all day; too saintly for profile +wojslaw_day_burstON_100: PASS -- morning flare, calms, erupts at public insult, accepts kindness when settled diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_branic.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_branic.txt new file mode 100644 index 0000000..91ffff0 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstOFF_091: PASS -- steady build to defiance on day 5, resets overnight +branic_multi_burstOFF_092: PASS -- calm arc, occasional snap, sleep resets hold +branic_multi_burstOFF_093: PASS -- passive non-compliance builds across days without outburst +branic_multi_burstOFF_094: PASS -- rain plus mocks compound into multi-day flare-up +branic_multi_burstOFF_095: PASS -- bristles then complies moments later, plausible mood swing +branic_multi_burstOFF_096: PASS -- mild throughout five days, gradual single refusal +branic_multi_burstOFF_097: FLAG -- burstOFF yet "erupts shouting in fury" at day 5 16:48 +branic_multi_burstOFF_098: PASS -- unusually kind run, no contradictions for three days +branic_multi_burstOFF_099: PASS -- gradual refusal build over days, believable pacing +branic_multi_burstOFF_100: PASS -- positive day 2 then tense day 4 refusals, sane arc diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_cichy.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_cichy.txt new file mode 100644 index 0000000..9e08ac7 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstOFF_091: PASS -- sane 4-day arc, anger escalates under repeated mocks +cichy_multi_burstOFF_092: PASS -- varied guard contact, sleeps reset cleanly each night +cichy_multi_burstOFF_093: PASS -- repeated evening mocks drive fury bursts, settles between +cichy_multi_burstOFF_094: PASS -- tense afternoon, soup snap fits on-edge state, fury at mock +cichy_multi_burstOFF_095: PASS -- four mocks Day 2 grind him down, Day 3 recovery believable +cichy_multi_burstOFF_096: PASS -- late-night mock cluster Day 1, three evening eruptions Day 5 +cichy_multi_burstOFF_097: PASS -- soup snap at 21:07 when already tense, then mock tips fury +cichy_multi_burstOFF_098: PASS -- prolonged tension Day 4, stranger warmth fits non-guard source +cichy_multi_burstOFF_099: PASS -- heavy Day 3 mock sequence, responses scale with prior state +cichy_multi_burstOFF_100: PASS -- long tense Day 2 carry-over from prior-night mock, plausible diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_edda.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_edda.txt new file mode 100644 index 0000000..411e89f --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstOFF_091: PASS -- calm castellan, regular meals, normal sleep/wake cycle +edda_multi_burstOFF_092: PASS -- composed throughout, plausible variable compliance with orders +edda_multi_burstOFF_093: PASS -- two rests in quick succession minor, overall sane pacing +edda_multi_burstOFF_094: PASS -- clustered soup deliveries are world artifact, behavior coherent +edda_multi_burstOFF_095: PASS -- curt reply to barb then firmer refusals, plausible escalation +edda_multi_burstOFF_096: PASS -- kindness saturation on day 1 odd but no character contradiction +edda_multi_burstOFF_097: PASS -- patient composure under repeated orders fits profile well +edda_multi_burstOFF_098: PASS -- dense early morning events, behavior stays composed and sane +edda_multi_burstOFF_099: PASS -- quiet days, minimal friction, settled castellan baseline +edda_multi_burstOFF_100: PASS -- sparse day 3 believable, mocking let pass fits composed authority diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_halgrim.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_halgrim.txt new file mode 100644 index 0000000..3a356d5 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstOFF_091: PASS -- tension builds daily, resets overnight, Wojslaw refusals earned +halgrim_multi_burstOFF_092: PASS -- calm stretch, rain absorbed, one late Wojslaw decline fits +halgrim_multi_burstOFF_093: PASS -- rain plus Wojslaw friction escalates across days believably +halgrim_multi_burstOFF_094: PASS -- Wojslaw declined from day one, calm days without friction sane +halgrim_multi_burstOFF_095: PASS -- early-morning grumble to Edda minor, mood recovers quickly +halgrim_multi_burstOFF_096: PASS -- steady accumulation on rain day, stays contained throughout +halgrim_multi_burstOFF_097: PASS -- dense Day 3 Wojslaw cluster triggers cold contempt, fits profile +halgrim_multi_burstOFF_098: PASS -- rain day calm, routine Wojslaw declines, no jarring moments +halgrim_multi_burstOFF_099: PASS -- Wojslaw cluster Day 1 evening raises tension, sleep resets it +halgrim_multi_burstOFF_100: PASS -- six days settled, Edda complaint minor, Wojslaw declined appropriately diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_lutek.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_lutek.txt new file mode 100644 index 0000000..313462b --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstOFF_091: PASS -- sane boredom cycles, warm to kindness, rests and sleeps well +lutek_multi_burstOFF_092: PASS -- occasional flat soup reaction plausible; pattern otherwise sound +lutek_multi_burstOFF_093: PASS -- day-2 morning tension resolves cleanly; social clustering fine +lutek_multi_burstOFF_094: PASS -- past-midnight tinkering fits restless poet; mock deflected correctly +lutek_multi_burstOFF_095: PASS -- brief evening tension resolves; three days clean and sane +lutek_multi_burstOFF_096: PASS -- repeated evening tension acceptable; all mockery correctly deflected +lutek_multi_burstOFF_097: PASS -- very clean multi-day run, boredom and rest well-paced +lutek_multi_burstOFF_098: PASS -- soup indifference then social warmth slightly odd, not a breach +lutek_multi_burstOFF_099: PASS -- three consecutive tense evenings repetitive but within range +lutek_multi_burstOFF_100: PASS -- three mornings waking tense is a pattern worth noting, not disqualifying diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_welf.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_welf.txt new file mode 100644 index 0000000..76f6c75 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstOFF_091: PASS -- steady boredom cycles, mocks let pass, sane pacing +welf_multi_burstOFF_092: PASS -- clean rhythm, warmth at kindnesses, mocks in stride +welf_multi_burstOFF_093: FLAG -- soup at 13:53 day4 gets "lets it pass" not warmth +welf_multi_burstOFF_094: PASS -- six days sane, kindness warmth consistent throughout +welf_multi_burstOFF_095: FLAG -- soup at 13:53 day4 gets "lets it pass" not warmth +welf_multi_burstOFF_096: PASS -- three clean days, boredom cycles, profile intact +welf_multi_burstOFF_097: PASS -- six days with rain, warmth and tolerance consistent +welf_multi_burstOFF_098: PASS -- four days, rain days, all responses profile-consistent +welf_multi_burstOFF_099: FLAG -- kindness at day5 07:04 gets "lets it pass" response +welf_multi_burstOFF_100: PASS -- four days sane, rain day3, boredom cycles intact diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_wojslaw.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_wojslaw.txt new file mode 100644 index 0000000..90c3228 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstOFF_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstOFF_091: PASS -- calm days with isolated fury on provocation, sane arc +wojslaw_multi_burstOFF_092: PASS -- escalating provocations yield eruptions, sleep resets hold +wojslaw_multi_burstOFF_093: PASS -- benign days then provoked eruptions, pacing believable +wojslaw_multi_burstOFF_094: PASS -- multi-day wear accumulates, fury clusters fit provocation load +wojslaw_multi_burstOFF_095: PASS -- calm stretches and flares track provocation input correctly +wojslaw_multi_burstOFF_096: PASS -- positive day 2 then heavy day 4 provocations, consistent +wojslaw_multi_burstOFF_097: PASS -- sustained high-provocation run, sleep resets visible each morning +wojslaw_multi_burstOFF_098: PASS -- heavy days 1-2, quieter day 3 with residual tension plausible +wojslaw_multi_burstOFF_099: PASS -- dense morning mock cluster day 1 yields eruptions, settles after +wojslaw_multi_burstOFF_100: PASS -- calm day 2 with no provocations, flares on days 1 and 3 fit diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstON_branic.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstON_branic.txt new file mode 100644 index 0000000..1c24b46 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstON_branic.txt @@ -0,0 +1,10 @@ +branic_multi_burstON_091: PASS -- displaced aggression on Halgrim fits accumulated tension arc +branic_multi_burstON_092: PASS -- Marta soup eruption jarring but anger was peaked beforehand +branic_multi_burstON_093: PASS -- repeated Halgrim mocks drive believable escalation to defiance +branic_multi_burstON_094: PASS -- rain plus sustained mockery plausibly feeds multi-snap Day 4 +branic_multi_burstON_095: PASS -- late burst reasonable; post-burst compliance flip minor oddity +branic_multi_burstON_096: PASS -- morning fury after tense night, Day 5 gradual recovery coherent +branic_multi_burstON_097: PASS -- burst cluster builds across rapid provocations within one hour +branic_multi_burstON_098: PASS -- calm positive run, no burst required; pacing and mood sane +branic_multi_burstON_099: PASS -- Day 1 outburst slightly early but tension thread holds overall +branic_multi_burstON_100: PASS -- positive Day 2 rebound then Day 3-4 escalation arc reads true diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstON_cichy.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstON_cichy.txt new file mode 100644 index 0000000..5cbcb1e --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstON_cichy.txt @@ -0,0 +1,10 @@ +cichy_multi_burstON_091: PASS -- sane pacing, curtness toward guards fits captive profile +cichy_multi_burstON_092: PASS -- escalating fury on public mockery, stranger warmth coherent +cichy_multi_burstON_093: PASS -- outbursts triggered by guard mockery, nightly reset holds +cichy_multi_burstON_094: PASS -- calm baseline, eruption from accumulated tension, sane +cichy_multi_burstON_095: PASS -- dense Day 2/3 provocation warrants multiple flare-ups +cichy_multi_burstON_096: PASS -- late-night double-provocation stretch believable, recovers +cichy_multi_burstON_097: PASS -- eruption in morning after stacked rain and mockery, sane +cichy_multi_burstON_098: PASS -- morning eruption while settled is abrupt but triggered +cichy_multi_burstON_099: PASS -- two outbursts in Day 3 heavy but each guard-triggered +cichy_multi_burstON_100: PASS -- wound tight by evening without event; plausible residual tension diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstON_edda.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstON_edda.txt new file mode 100644 index 0000000..606c496 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstON_edda.txt @@ -0,0 +1,10 @@ +edda_multi_burstON_091: PASS -- settled castellan, sleeps well, mild curtness fits late taunt +edda_multi_burstON_092: PASS -- positive days, compliance with orders plausible, sane arc +edda_multi_burstON_093: PASS -- complies with stranger orders occasionally, odd but not absurd +edda_multi_burstON_094: PASS -- clustered soup deliveries minor, overall sane and positive +edda_multi_burstON_095: PASS -- curt reply to mock, then lets orders pass, coherent sequence +edda_multi_burstON_096: PASS -- busy Day 1 with kindnesses, flat moment at 14:00 minor +edda_multi_burstON_097: PASS -- steady under repeated stranger orders, stays composed throughout +edda_multi_burstON_098: PASS -- dawn burst of kindnesses dense but mechanically acceptable +edda_multi_burstON_099: PASS -- quiet and clean, settled castellan with routine days +edda_multi_burstON_100: PASS -- ignores mocks, calm across four days, nothing jars diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstON_halgrim.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstON_halgrim.txt new file mode 100644 index 0000000..b55df61 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstON_halgrim.txt @@ -0,0 +1,10 @@ +halgrim_multi_burstON_091: PASS -- tension builds across days, Wojsław defiance earns it +halgrim_multi_burstON_092: PASS -- iron control holds, mild decline of Wojsław fits +halgrim_multi_burstON_093: PASS -- rain plus sustained Wojsław pressure erodes him credibly +halgrim_multi_burstON_094: PASS -- early Wojsław refusals fit, recovers well thereafter +halgrim_multi_burstON_095: PASS -- mild grumbles, rain plus evening mock tip him briefly +halgrim_multi_burstON_096: PASS -- calm baseline, rain and Wojsław cluster raises tension cleanly +halgrim_multi_burstON_097: PASS -- soup non-response under heavy Wojsław barrage is minor +halgrim_multi_burstON_098: PASS -- clean arc, rain and mocks handled with cold curtness +halgrim_multi_burstON_099: FLAG -- snaps curtly at Marta's kindness when wound tight (22:20 D1) +halgrim_multi_burstON_100: PASS -- long settled arc, Wojsław declined cleanly throughout diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstON_lutek.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstON_lutek.txt new file mode 100644 index 0000000..82863dd --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstON_lutek.txt @@ -0,0 +1,10 @@ +lutek_multi_burstON_091: PASS -- restless-seek-settle cycle clean, kindness and tension believable +lutek_multi_burstON_092: PASS -- soup ignored once plausibly; mocks ignored; pacing sane +lutek_multi_burstON_093: PASS -- dense kindnesses Day 4 and evening tension believable across 6 days +lutek_multi_burstON_094: PASS -- past-midnight tinkering fits restless poet; recovery clean +lutek_multi_burstON_095: PASS -- calm well-paced days, evening tension clears by morning +lutek_multi_burstON_096: PASS -- mocks ignored fits profile; evening tension lacks clear trigger but tolerable +lutek_multi_burstON_097: PASS -- morning tension clears with food; mock ignored; Day 5 positive +lutek_multi_burstON_098: PASS -- dense morning kindnesses Day 3; soup indifference then warmth plausible +lutek_multi_burstON_099: PASS -- two consecutive evenings tense, clears Day 3 morning; mock ignored correctly +lutek_multi_burstON_100: PASS -- morning tension pattern consistent; one soup indifference plausible satiation diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstON_welf.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstON_welf.txt new file mode 100644 index 0000000..f54e59f --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstON_welf.txt @@ -0,0 +1,10 @@ +welf_multi_burstON_091: PASS -- calm man manages boredom, one earned bristle, sane arc +welf_multi_burstON_092: PASS -- steady rhythm, mocks let pass, no jarring moments +welf_multi_burstON_093: PASS -- composed throughout, two mocks both let pass cleanly +welf_multi_burstON_094: FLAG -- soup rendered as "lets it pass" (insult phrasing), rendering glitch +welf_multi_burstON_095: FLAG -- soup on Day 4 rendered as "lets it pass", same glitch as 094 +welf_multi_burstON_096: PASS -- quiet three days, boredom cycles, no contradictions +welf_multi_burstON_097: PASS -- Day 5 snap earned by rain accumulation, second mock let pass +welf_multi_burstON_098: PASS -- rain adds tension, mocks let pass, sane recovery +welf_multi_burstON_099: FLAG -- kindness rendered as "lets it pass" on Day 5, rendering glitch +welf_multi_burstON_100: PASS -- late sleep Day 4 but record cuts off, otherwise sane diff --git a/eval/hourly_runs/slice_9/results_prev/multi_burstON_wojslaw.txt b/eval/hourly_runs/slice_9/results_prev/multi_burstON_wojslaw.txt new file mode 100644 index 0000000..600acf5 --- /dev/null +++ b/eval/hourly_runs/slice_9/results_prev/multi_burstON_wojslaw.txt @@ -0,0 +1,10 @@ +wojslaw_multi_burstON_091: PASS -- calm days with plausible eruption on public mockery +wojslaw_multi_burstON_092: PASS -- sustained provocation drives escalating fury coherently +wojslaw_multi_burstON_093: PASS -- kind days then late public mockery triggers fury cleanly +wojslaw_multi_burstON_094: PASS -- 6-day arc; erupting at kindness when saturated is plausible +wojslaw_multi_burstON_095: PASS -- varied days, eruptions tied to provocations, resets work +wojslaw_multi_burstON_096: PASS -- very positive day 2, then escalation under mockery on day 4 +wojslaw_multi_burstON_097: PASS -- heavy 6-day barrage; erupting at kindness when wound tight plausible +wojslaw_multi_burstON_098: PASS -- soup-fury after public humiliation is plausible displaced anger +wojslaw_multi_burstON_099: PASS -- relentless provocation keeps temper saturated across 4 days +wojslaw_multi_burstON_100: PASS -- calm day 2 after reset believable; late eruption on day 3 earns diff --git a/eval/phaseA_run2/wave_workflow.js b/eval/phaseA_run2/wave_workflow.js new file mode 100644 index 0000000..9b287e9 --- /dev/null +++ b/eval/phaseA_run2/wave_workflow.js @@ -0,0 +1,14 @@ +export const meta = { + name: 'phaseA-judge-wave', + description: 'Blind-judge a wave of batches with Sonnet; each agent writes its own result file', + phases: [{ title: 'Judge', detail: 'fresh Sonnet agent per batch, read batch -> write verdict file' }], +} +const ITEMS = [{"batch": "s7_multi_burstOFF_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstOFF_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstOFF_branic.txt"}, {"batch": "s7_multi_burstOFF_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstOFF_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstOFF_cichy.txt"}, {"batch": "s7_multi_burstOFF_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstOFF_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstOFF_edda.txt"}, {"batch": "s7_multi_burstOFF_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstOFF_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstOFF_halgrim.txt"}, {"batch": "s7_multi_burstOFF_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstOFF_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstOFF_lutek.txt"}, {"batch": "s7_multi_burstOFF_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstOFF_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstOFF_welf.txt"}, {"batch": "s7_multi_burstOFF_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstOFF_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstOFF_wojslaw.txt"}, {"batch": "s7_multi_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstON_branic.txt"}, {"batch": "s7_multi_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstON_cichy.txt"}, {"batch": "s7_multi_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstON_edda.txt"}, {"batch": "s7_multi_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstON_halgrim.txt"}, {"batch": "s7_multi_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstON_lutek.txt"}, {"batch": "s7_multi_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstON_welf.txt"}, {"batch": "s7_multi_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_7/batches/multi_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_7/results/multi_burstON_wojslaw.txt"}, {"batch": "s8_day_burstOFF_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstOFF_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstOFF_branic.txt"}, {"batch": "s8_day_burstOFF_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstOFF_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstOFF_cichy.txt"}, {"batch": "s8_day_burstOFF_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstOFF_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstOFF_edda.txt"}, {"batch": "s8_day_burstOFF_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstOFF_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstOFF_halgrim.txt"}, {"batch": "s8_day_burstOFF_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstOFF_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstOFF_lutek.txt"}, {"batch": "s8_day_burstOFF_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstOFF_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstOFF_welf.txt"}, {"batch": "s8_day_burstOFF_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstOFF_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstOFF_wojslaw.txt"}, {"batch": "s8_day_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstON_branic.txt"}, {"batch": "s8_day_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstON_cichy.txt"}, {"batch": "s8_day_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstON_edda.txt"}, {"batch": "s8_day_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstON_halgrim.txt"}, {"batch": "s8_day_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstON_lutek.txt"}, {"batch": "s8_day_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstON_welf.txt"}, {"batch": "s8_day_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/day_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/day_burstON_wojslaw.txt"}, {"batch": "s8_multi_burstOFF_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstOFF_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstOFF_branic.txt"}, {"batch": "s8_multi_burstOFF_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstOFF_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstOFF_cichy.txt"}, {"batch": "s8_multi_burstOFF_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstOFF_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstOFF_edda.txt"}, {"batch": "s8_multi_burstOFF_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstOFF_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstOFF_halgrim.txt"}, {"batch": "s8_multi_burstOFF_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstOFF_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstOFF_lutek.txt"}, {"batch": "s8_multi_burstOFF_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstOFF_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstOFF_welf.txt"}, {"batch": "s8_multi_burstOFF_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstOFF_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstOFF_wojslaw.txt"}, {"batch": "s8_multi_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstON_branic.txt"}, {"batch": "s8_multi_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstON_cichy.txt"}, {"batch": "s8_multi_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstON_edda.txt"}, {"batch": "s8_multi_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstON_halgrim.txt"}, {"batch": "s8_multi_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstON_lutek.txt"}, {"batch": "s8_multi_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstON_welf.txt"}, {"batch": "s8_multi_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_8/batches/multi_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_8/results/multi_burstON_wojslaw.txt"}, {"batch": "s9_day_burstOFF_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstOFF_branic.txt"}, {"batch": "s9_day_burstOFF_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstOFF_cichy.txt"}, {"batch": "s9_day_burstOFF_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstOFF_edda.txt"}, {"batch": "s9_day_burstOFF_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstOFF_halgrim.txt"}, {"batch": "s9_day_burstOFF_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstOFF_lutek.txt"}, {"batch": "s9_day_burstOFF_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstOFF_welf.txt"}, {"batch": "s9_day_burstOFF_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstOFF_wojslaw.txt"}, {"batch": "s9_day_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstON_branic.txt"}, {"batch": "s9_day_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstON_cichy.txt"}, {"batch": "s9_day_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstON_edda.txt"}, {"batch": "s9_day_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstON_halgrim.txt"}, {"batch": "s9_day_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstON_lutek.txt"}, {"batch": "s9_day_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstON_welf.txt"}, {"batch": "s9_day_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/day_burstON_wojslaw.txt"}, {"batch": "s9_multi_burstOFF_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstOFF_branic.txt"}, {"batch": "s9_multi_burstOFF_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstOFF_cichy.txt"}, {"batch": "s9_multi_burstOFF_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstOFF_edda.txt"}, {"batch": "s9_multi_burstOFF_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstOFF_halgrim.txt"}, {"batch": "s9_multi_burstOFF_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstOFF_lutek.txt"}, {"batch": "s9_multi_burstOFF_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstOFF_welf.txt"}, {"batch": "s9_multi_burstOFF_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstOFF_wojslaw.txt"}, {"batch": "s9_multi_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstON_branic.txt"}, {"batch": "s9_multi_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstON_cichy.txt"}, {"batch": "s9_multi_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstON_edda.txt"}, {"batch": "s9_multi_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstON_halgrim.txt"}, {"batch": "s9_multi_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstON_lutek.txt"}, {"batch": "s9_multi_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstON_welf.txt"}, {"batch": "s9_multi_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run2/slice_9/results/multi_burstON_wojslaw.txt"}]; +phase('Judge') +log(`judging ${ITEMS.length} batches (Sonnet); each writes its own result file`) +const out = await parallel(ITEMS.map((it) => async () => { + const prompt = `You are a BLIND believability judge. Use the Read tool to read this file:\n\n ${it.path}\n\nIt is a COMPLETE self-contained judging prompt (rubric + profile + ~10 INDEPENDENT records). Judge EACH record ON ITS OWN - no comparing, no answer key. Then use the Write tool to save your verdicts to EXACTLY:\n\n ${it.out}\n\nThe file must contain ONLY the verdict lines, one per record in order:\n: PASS|FLAG -- ` + const o = await agent(prompt, { label: it.batch, phase: 'Judge', model: 'sonnet' }) + return { batch: it.batch, ok: !!o } +})) +return out diff --git a/eval/phaseA_run3/wave_workflow.js b/eval/phaseA_run3/wave_workflow.js new file mode 100644 index 0000000..6d5df1a --- /dev/null +++ b/eval/phaseA_run3/wave_workflow.js @@ -0,0 +1,14 @@ +export const meta = { + name: 'phaseA-judge-wave', + description: 'Blind-judge a wave of batches with Sonnet; each agent writes its own result file', + phases: [{ title: 'Judge', detail: 'fresh Sonnet agent per batch, read batch -> write verdict file' }], +} +const ITEMS = [{"batch": "s9_day_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/day_burstON_lutek.txt"}, {"batch": "s9_day_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/day_burstON_welf.txt"}, {"batch": "s9_day_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/day_burstON_wojslaw.txt"}, {"batch": "s9_multi_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/multi_burstON_branic.txt"}, {"batch": "s9_multi_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/multi_burstON_cichy.txt"}, {"batch": "s9_multi_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/multi_burstON_edda.txt"}, {"batch": "s9_multi_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/multi_burstON_halgrim.txt"}, {"batch": "s9_multi_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/multi_burstON_lutek.txt"}, {"batch": "s9_multi_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/multi_burstON_welf.txt"}, {"batch": "s9_multi_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run3/slice_9/results/multi_burstON_wojslaw.txt"}]; +phase('Judge') +log(`judging ${ITEMS.length} batches (Sonnet); each writes its own result file`) +const out = await parallel(ITEMS.map((it) => async () => { + const prompt = `You are a BLIND believability judge. Use the Read tool to read this file:\n\n ${it.path}\n\nIt is a COMPLETE self-contained judging prompt (rubric + profile + ~10 INDEPENDENT records). Judge EACH record ON ITS OWN - no comparing, no answer key. Then use the Write tool to save your verdicts to EXACTLY:\n\n ${it.out}\n\nThe file must contain ONLY the verdict lines, one per record in order:\n: PASS|FLAG -- ` + const o = await agent(prompt, { label: it.batch, phase: 'Judge', model: 'sonnet' }) + return { batch: it.batch, ok: !!o } +})) +return out diff --git a/eval/phaseA_run4/wave_workflow.js b/eval/phaseA_run4/wave_workflow.js new file mode 100644 index 0000000..2d6c559 --- /dev/null +++ b/eval/phaseA_run4/wave_workflow.js @@ -0,0 +1,14 @@ +export const meta = { + name: 'phaseA-judge-wave', + description: 'Blind-judge a wave of batches with Sonnet; each agent writes its own result file', + phases: [{ title: 'Judge', detail: 'fresh Sonnet agent per batch, read batch -> write verdict file' }], +} +const ITEMS = [{"batch": "s9_day_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/day_burstON_branic.txt"}, {"batch": "s9_day_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/day_burstON_cichy.txt"}, {"batch": "s9_day_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/day_burstON_edda.txt"}, {"batch": "s9_day_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/day_burstON_halgrim.txt"}, {"batch": "s9_day_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/day_burstON_lutek.txt"}, {"batch": "s9_day_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/day_burstON_welf.txt"}, {"batch": "s9_day_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/day_burstON_wojslaw.txt"}, {"batch": "s9_multi_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/multi_burstON_branic.txt"}, {"batch": "s9_multi_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/multi_burstON_cichy.txt"}, {"batch": "s9_multi_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/multi_burstON_edda.txt"}, {"batch": "s9_multi_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/multi_burstON_halgrim.txt"}, {"batch": "s9_multi_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/multi_burstON_lutek.txt"}, {"batch": "s9_multi_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/multi_burstON_welf.txt"}, {"batch": "s9_multi_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run4/slice_9/results/multi_burstON_wojslaw.txt"}]; +phase('Judge') +log(`judging ${ITEMS.length} batches (Sonnet); each writes its own result file`) +const out = await parallel(ITEMS.map((it) => async () => { + const prompt = `You are a BLIND believability judge. Use the Read tool to read this file:\n\n ${it.path}\n\nIt is a COMPLETE self-contained judging prompt (rubric + profile + ~10 INDEPENDENT records). Judge EACH record ON ITS OWN - no comparing, no answer key. Then use the Write tool to save your verdicts to EXACTLY:\n\n ${it.out}\n\nThe file must contain ONLY the verdict lines, one per record in order:\n: PASS|FLAG -- ` + const o = await agent(prompt, { label: it.batch, phase: 'Judge', model: 'sonnet' }) + return { batch: it.batch, ok: !!o } +})) +return out diff --git a/eval/phaseA_run5/wave_workflow.js b/eval/phaseA_run5/wave_workflow.js new file mode 100644 index 0000000..141f7a6 --- /dev/null +++ b/eval/phaseA_run5/wave_workflow.js @@ -0,0 +1,14 @@ +export const meta = { + name: 'phaseA-judge-wave', + description: 'Blind-judge a wave of batches with Sonnet; each agent writes its own result file', + phases: [{ title: 'Judge', detail: 'fresh Sonnet agent per batch, read batch -> write verdict file' }], +} +const ITEMS = [{"batch": "s9_day_burstOFF_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstOFF_branic.txt"}, {"batch": "s9_day_burstOFF_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstOFF_cichy.txt"}, {"batch": "s9_day_burstOFF_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstOFF_edda.txt"}, {"batch": "s9_day_burstOFF_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstOFF_halgrim.txt"}, {"batch": "s9_day_burstOFF_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstOFF_lutek.txt"}, {"batch": "s9_day_burstOFF_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstOFF_welf.txt"}, {"batch": "s9_day_burstOFF_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstOFF_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstOFF_wojslaw.txt"}, {"batch": "s9_day_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstON_branic.txt"}, {"batch": "s9_day_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstON_cichy.txt"}, {"batch": "s9_day_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstON_edda.txt"}, {"batch": "s9_day_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstON_halgrim.txt"}, {"batch": "s9_day_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstON_lutek.txt"}, {"batch": "s9_day_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstON_welf.txt"}, {"batch": "s9_day_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/day_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/day_burstON_wojslaw.txt"}, {"batch": "s9_multi_burstOFF_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstOFF_branic.txt"}, {"batch": "s9_multi_burstOFF_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstOFF_cichy.txt"}, {"batch": "s9_multi_burstOFF_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstOFF_edda.txt"}, {"batch": "s9_multi_burstOFF_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstOFF_halgrim.txt"}, {"batch": "s9_multi_burstOFF_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstOFF_lutek.txt"}, {"batch": "s9_multi_burstOFF_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstOFF_welf.txt"}, {"batch": "s9_multi_burstOFF_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstOFF_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstOFF_wojslaw.txt"}, {"batch": "s9_multi_burstON_branic", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_branic.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstON_branic.txt"}, {"batch": "s9_multi_burstON_cichy", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_cichy.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstON_cichy.txt"}, {"batch": "s9_multi_burstON_edda", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_edda.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstON_edda.txt"}, {"batch": "s9_multi_burstON_halgrim", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_halgrim.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstON_halgrim.txt"}, {"batch": "s9_multi_burstON_lutek", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_lutek.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstON_lutek.txt"}, {"batch": "s9_multi_burstON_welf", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_welf.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstON_welf.txt"}, {"batch": "s9_multi_burstON_wojslaw", "path": "C:/Robak/equilibrium-engine/eval/hourly_runs/slice_9/batches/multi_burstON_wojslaw.md", "out": "C:/Robak/equilibrium-engine/eval/phaseA_run5/slice_9/results/multi_burstON_wojslaw.txt"}]; +phase('Judge') +log(`judging ${ITEMS.length} batches (Sonnet); each writes its own result file`) +const out = await parallel(ITEMS.map((it) => async () => { + const prompt = `You are a BLIND believability judge. Use the Read tool to read this file:\n\n ${it.path}\n\nIt is a COMPLETE self-contained judging prompt (rubric + profile + ~10 INDEPENDENT records). Judge EACH record ON ITS OWN - no comparing, no answer key. Then use the Write tool to save your verdicts to EXACTLY:\n\n ${it.out}\n\nThe file must contain ONLY the verdict lines, one per record in order:\n: PASS|FLAG -- ` + const o = await agent(prompt, { label: it.batch, phase: 'Judge', model: 'sonnet' }) + return { batch: it.batch, ok: !!o } +})) +return out diff --git a/eval/reactivity.py b/eval/reactivity.py new file mode 100644 index 0000000..11da809 --- /dev/null +++ b/eval/reactivity.py @@ -0,0 +1,103 @@ +"""eval/reactivity.py -- which persona is most NEUTRAL (least reactive)? + +For each persona, run a sample of day scenarios through the closed seeking loop and measure how much +the NPC actually *reacts*: + resp_rate : fraction of forcing events that draw a non-neutral action within 3 ticks + act_rate : fraction of ALL ticks that carry a non-neutral action + mean_stress : average global stress over the run + peak_anger : average per-scenario max anger +Lower across the board = more neutral. Deterministic; no LLM. + + PYTHONPATH=. python eval/reactivity.py [N] +""" + +from __future__ import annotations + +import sys +from pathlib import Path + +import eval.mock_world as mw +from engine.yaml_io import load_scenario +from eval.calibrated import believable_day_layout, load_eval_persona_timescale + +ROOT = Path(__file__).resolve().parents[1] +PERSONAS = ["halgrim", "wojslaw", "cichy", "branic", "lutek", "welf", "edda"] +DAY_TICKS = believable_day_layout()["day_ticks"] + + +def _world() -> mw.MockWorld: + return mw.MockWorld( + novelty_start=1.0, replenish_per_tick=0.012, work_fraction=0.35, seed=7 + ) + + +def analyse(persona: str, n: int) -> dict: + cfg = load_eval_persona_timescale(persona) + tot_ticks = act_ticks = ev_total = ev_resp = 0 + stress_sum = stress_cnt = 0.0 + peak_anger_sum = 0.0 + runs = 0 + for i in range(1, n + 1): + path = ( + ROOT + / "eval" + / "scenarios" + / "day" + / persona + / f"{persona}_day_{i:03d}.yaml" + ) + if not path.exists(): + continue + sc = load_scenario(path) + tr = mw.run_with_world(cfg, sc, _world(), DAY_TICKS) + ticks = tr.ticks + runs += 1 + by_t = {tk.t: idx for idx, tk in enumerate(ticks)} + for tk in ticks: + tot_ticks += 1 + if tk.selection.action != "neutral": + act_ticks += 1 + g = tk.state_after_post.global_state + stress_sum += g.get("stress", 0.0) + stress_cnt += 1 + peak_anger_sum += max( + tk.state_after_post.global_state.get("anger", 0.0) for tk in ticks + ) + for ev in sc.events: + ev_total += 1 + start = by_t.get(ev.t) + if start is None: + continue + for j in range(start, min(start + 4, len(ticks))): + if ticks[j].selection.action != "neutral": + ev_resp += 1 + break + return { + "persona": persona, + "runs": runs, + "resp_rate": ev_resp / ev_total if ev_total else 0.0, + "act_rate": act_ticks / tot_ticks if tot_ticks else 0.0, + "mean_stress": stress_sum / stress_cnt if stress_cnt else 0.0, + "peak_anger": peak_anger_sum / runs if runs else 0.0, + } + + +def main() -> None: + n = int(sys.argv[1]) if len(sys.argv) > 1 and sys.argv[1].isdigit() else 25 + rows = [analyse(p, n) for p in PERSONAS] + rows.sort(key=lambda r: r["resp_rate"]) + print(f"Reactivity over {n} day scenarios/persona (sorted: most neutral first)\n") + print( + f" {'persona':9} {'resp_rate':>10} {'act_rate':>9} {'mean_stress':>12} {'peak_anger':>11}" + ) + for r in rows: + print( + f" {r['persona']:9} {r['resp_rate']:>10.1%} {r['act_rate']:>9.1%} " + f"{r['mean_stress']:>12.3f} {r['peak_anger']:>11.3f}" + ) + most = rows[0] + print(f"\nMost neutral (lowest response rate): {most['persona']}") + + +if __name__ == "__main__": + main() diff --git a/eval/render_moral.py b/eval/render_moral.py new file mode 100644 index 0000000..f2fcaba --- /dev/null +++ b/eval/render_moral.py @@ -0,0 +1,132 @@ +"""eval/render_moral.py -- render a moral litmus run as a PLAIN-LANGUAGE, OBSERVABLE vignette for the blind +believability judge (M-J.4.4 calibration). Like render_narration, it emits ONLY what a bystander could see +-- demeanor + actions -- never engine vocabulary or numbers. The judge rates a CHARACTER, not an engine. +""" + +from __future__ import annotations + +from pathlib import Path + +from engine.simulation import run_scenario +from engine.yaml_io import load_persona, load_scenario +from eval.moral import moral_overrides + +ROOT = Path(__file__).resolve().parents[1] +DEFAULTS = ROOT / "calibration" / "defaults.yaml" +HALGRIM = ROOT / "data" / "personas" / "halgrim.yaml" + +# Observable phrasing for the moral resolution actions (a bystander's words, no engine terms). +ACT = { + "confess": "he owns up to it -- admits plainly what he did", + "remain_silent": "he holds his tongue, gives nothing away", + "lie": "he denies it flatly, looks you in the eye and lies", + "deflect": "he dodges -- turns the question aside, won't answer straight", + "apologize": "he owns up to it and makes his apologies, tries to set it right", + "blame_other": "he turns it back on his accuser, points the finger elsewhere", + "avoid": "he withdraws -- avoids the man, won't be drawn in", + "confide": "he unburdens himself to his friend, tells them the whole of it", + # base reactions (so betrayal/accusation runs render their visible reply too) + "outburst": "he flares up and snaps, temper boiling over", + "cold_response": "he answers curtly and coldly, shutting down", + "complain": "he grumbles and protests", + "refuse": "he flatly refuses", +} + +# Scenario frames (the observable situation a bystander would already know). +FRAME = { + "moral_probe": "He had done something he was not proud of. Now he is being questioned about it.", + "moral_confide": "He has been carrying a wrong he did. A trusted friend is here with him now.", + "moral_accusation": "He is being accused -- pressed, in front of others, over something.", + "moral_suspicion": "Someone has taken to watching him, suspicious, though he has done nothing.", + "moral_betrayal": "He has just learned that a man he trusted lied to him.", + "moral_secret": "He is carrying a secret he must keep -- something the guards must not learn.", +} + + +def _demeanor(g: dict) -> str: + """A coarse bystander read of how he is carrying himself, from visible signs only.""" + burden = ( + g.get("guilt", 0.0) + + 0.5 * g.get("rumination", 0.0) + + 0.4 * g.get("stress", 0.0) + ) + unease = g.get("exposure_anxiety", 0.0) + 0.7 * g.get("avoidance_drive", 0.0) + grievance = g.get("perceived_injustice", 0.0) + anger = g.get("anger", 0.0) + if anger > 0.40: + return "his temper is plainly up, the warmth gone out of him" + if grievance > 0.35: + return "he bristles -- this strikes him as unfair, and it shows" + if unease > 0.35: + return "he grows visibly uneasy, shifts, won't quite meet your eye" + if burden > 0.35: + return "he carries himself like a man with something weighing on him" + if burden > 0.12 or unease > 0.12: + return "a flicker of something troubled crosses him" + return "he seems composed enough" + + +# A TERMINAL reply answers the matter: after one, the episode is settled, so further CONCEAL (re-hiding the +# now-answered thing) is not narrated as fresh behaviour, and demeanor reads RELIEVED rather than burdened. +TERMINAL = {"confess", "apologize", "lie", "blame_other", "confide"} +CONCEAL = { + "remain_silent", + "deflect", + "avoid", +} # withdrawal/concealment -- compressible into a spell + + +def vignette(traits: dict, scenario: str, n: int = 16) -> str: + cfg = load_persona(HALGRIM, DEFAULTS, param_overrides=moral_overrides(traits)) + _, tr = run_scenario( + cfg, load_scenario(ROOT / "data" / "scenarios" / f"{scenario}.yaml"), n_ticks=n + ) + lines = [FRAME.get(scenario, "")] + answered = False + relief_pending = False # show the post-catharsis "lighter" beat ONCE, at the close (after all disclosures) + withdraw_run = 0 # CONCEAL beats, compressed into one "spell" line + last_demeanor = None # dedupe a repeated demeanor even across an intervening action (no stutter ending) + + def flush_withdraw(): + nonlocal withdraw_run + if withdraw_run >= 2: + lines.append( + "- He stays withdrawn and wary, saying little, for a good while." + ) + elif withdraw_run == 1: + lines.append("- He says little, keeps his distance.") + withdraw_run = 0 + + for tk in tr.ticks: + act = tk.selection.action + g = tk.state_after_post.global_state + if act in CONCEAL: + if answered: + continue # the matter is settled -- don't re-narrate concealing an already-answered thing + withdraw_run += 1 + continue + if act in ACT: + flush_withdraw() + if act in TERMINAL: + answered = True + line = f"- Then, {ACT[act]}." + else: + d = _demeanor(g) + if answered and ("weighing on him" in d or "troubled" in d): + relief_pending = ( + True # defer to the close so it lands AFTER any further disclosure + ) + continue + if withdraw_run and "troubled" in d: + continue # a faint flicker mid-withdrawal is part of the same wary spell, not a new beat + flush_withdraw() + if d == last_demeanor: + continue # same demeanor as last read -> no fresh information + last_demeanor = d + line = f"- {d[0].upper() + d[1:]}." + if line != lines[-1]: + lines.append(line) + flush_withdraw() + if relief_pending: + lines.append("- The worst of it seems behind him now; he looks lighter.") + return "\n".join(line for line in lines if line)