Repository navigation
M-J A3 -- minor/serious guilt split via Secret.moral_weight - #16
Merged
Merged
Conversation
…eight The blind multi-day judge validated serious-guilt ~72h but flagged 18h as "too tidy for a serious wrong" -- yet the engine has one guilt state/half-life. Rather than two states, an ACTIVE secret now RE-INJECTS guilt each tick by secret_weight_to_guilt * moral_weight * salience (engine/simulation.py _update_secrets, the post_effects phase): a SERIOUS unconfessed wrong (high moral_weight) keeps guilt alive against decay -> it lingers (the 72h feel) while the base half-life stays minor-appropriate (18h); a minor wrong fades; confession/exposure inactivates the secret -> the drip stops -> relief. This reuses the Secret lifecycle and gives the spec section 11 minor/serious split with no new state. - moral_overlay: ledger_params.secret_weight_to_guilt (placeholder). - tests/test_moral_guilt_weight.py (3): serious secret sustains guilt while minor fades; exposure stops the drip (relief); opt-in (legacy/no-overlay inert). Full suite 399 passed +4 skipped; deterministic pre-filter still 18/81 (no litmus broken -- re-ran after the magnitude change, per the calibration lesson); goldens byte-identical; ruff clean. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016rWVvvhWijSuLMEhzjiUhu
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
A3 — minor/serious guilt split (spec §11)
The blind multi-day judge validated serious-guilt persistence at ~72h but flagged 18h as "too tidy for a serious wrong" — yet the engine has ONE guilt state with ONE half-life. Rather than add a second state, an ACTIVE secret re-injects guilt each tick by
secret_weight_to_guilt · moral_weight · salience(in_update_secrets, post_effects phase):moral_weight) keeps guilt alive against decay → it lingers (the 72h feel), while the base half-life stays minor-appropriate (18h);This reuses the Secret lifecycle (no new state) and realises the spec §11 split via the authored
moral_weight.Tests
tests/test_moral_guilt_weight.py(3): serious secret sustains guilt while minor fades; exposure stops the drip; opt-in (legacy/no-overlay inert).Status
Full suite 399 passed, 4 skipped; deterministic pre-filter still 18/81 (re-ran after the magnitude change, per the calibration lesson — no litmus broken); goldens byte-identical; ruff clean.
🤖 Generated with Claude Code