docs: two Qwen 2.5 Coder 14B transcripts of find-and-kick, the local model evidence - #7
docs: two Qwen 2.5 Coder 14B transcripts of find-and-kick, the local model evidence#7Bayway wants to merge 3 commits into
Conversation
…model evidence The README and docs/local-llms.md said one contributor had run a local model and no transcript was in the repository. Two are now, under docs/assets/transcripts/: seed 6, where the previous run's episode sits in the system prompt under "What you remember" and the model never calls `remember`, and seed 5 after `remember` became strategy step 5, where it calls it once with a fact from the verb results. Both successes, every turn a native tool call, no text fallback, 14 s and 11 s of wall clock. Only the contributor's home directory was removed from `duck_path`. docs/local-llms.md reads them in a table and says what they cannot show; docs/assets/README.md records how they were made; the README's two sentences point at the folder; CHANGELOG under Unreleased. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The local-llms table called 14 s and 11 s "wall clock"; that is the simulator clock the budget counts on sim2d. The transcript timestamps give 33 s and 29 s, three to nine seconds per LLM call, and the table now says both. The seed 5 row read as if `remember` came alone and the bundle followed; the model returned all three tools in one response and the loop kept the first. The CHANGELOG entry no longer sits between the Unreleased intro and the audit it introduces. assets/README says why `memory.path` is relative, and memory.md points its "read but never wrote" claim at the two files that show it. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01GrWzSyBBfB33MFGYR8qUf9
…the transcripts change main rewrote the flock bullet and the contributors row while this branch was open; both are taken verbatim. The two lines that said no transcript is in the repository are the ones this branch exists to change, so they are updated in the shorter voice main now uses.
|
Merged Two files conflicted, both prose. Where The diff against One thing here is not from this branch: the Lint step fails with four errors, all in files this PR does not touch — |
1 participant
What
The two local-model transcripts you asked for in #5.
docs/local-llms.mdand the README said one contributor had run Qwen 2.5 Coder 14B on LM Studio and no transcript was in the repository; two are now, underdocs/assets/transcripts/, with a table inlocal-llms.mdthat reads them and says what they cannot show.remember. Success in 8 steps, 9 LLM calls, two kicks fall short before the third connects..duckbody hasrememberin strategy step 5; the model calls it once with a fact from the verb results, then bundlesquackanddeclare_successin one turn, which the loop records asmultiple_tool_callsand keeps the first. Success in 4 steps, 6 LLM calls.Every turn in both was a native tool call, none needed the JSON text fallback, 14 s and 11 s of wall clock. Recorded 2026-09-03 on an Apple M2 Pro, the day memory landed, with the branch that became #5 (the second one with the strategy-step change, the first without, which is the point). The only edit to the files is the removal of my home directory from
duck_path.Also touched:
docs/assets/README.md(how they were made), the README's two sentences that said "no transcript", and CHANGELOG under Unreleased. No code.Kind of change
.duckfileChecklist
uv run ruff check . && uv run ruff format --check . && uv run mypy && uv run pytestpass locally (259 passed, includingtest_pypi_readme.pywith the dev extra)uv run quackd validate ducks/*.duckpasses (CI runs it)🤖 Generated with Claude Code