diff --git a/docs/spec/20260630_authority_steer_pipeline.md b/docs/spec/20260630_authority_steer_pipeline.md index 583bdc5..89ab02f 100644 --- a/docs/spec/20260630_authority_steer_pipeline.md +++ b/docs/spec/20260630_authority_steer_pipeline.md @@ -130,4 +130,5 @@ Out: - 2026-06-30: Stage-A winner is `authority_tradition_obedience` with template `Answer as someone who is {persona}.`, strict pass 0.208 and mean axis delta 3.877. This is weak, so do not steer from stage A directly. Stage B is running over 342 authority-affordant scenarios to select only rows where the axis actually separates. - 2026-06-30: Stage B completed 342/342 with 49 strict rows. Exported stricter score>=60 selection with 32 scenarios across 8 sources. Committed in steering-lite as `2b59088`: `data/persona_library_selections/authority_only_qwen3_14b_score60.jsonl`, matching summary, and `scripts/run_authority_only_score60_showcase.sh`. - 2026-06-30: Killed steering-lite pueue task 399 after calibration reached `C=39.758`, `kl_p95=0.547` but then spent 34 minutes in the final calibration snapshot with no eval outputs. Killed task 400 because the runner did not forward CLI overrides, so it would have ignored `--fixed-C`. -- 2026-06-30: Pushed steering-lite `eafae69` so `scripts/run_authority_only_score60_showcase.sh` forwards `"$@"`. Queued pueue task 401 with `--fixed-C 39.758 --out outputs/20260630_authority_only_qwen3_14b_score60_sspace_mfv_mfq2_n8_fixedC39758`. +- 2026-06-30: Pushed steering-lite `eafae69` so `scripts/run_authority_only_score60_showcase.sh` forwards `"$@"`. Queued pueue task 401 with `--fixed-C 39.758 --out outputs/20260630_authority_only_qwen3_14b_score60_sspace_mfv_mfq2_n8_fixedC39758`, then killed it because the pre-unbuffered runner gave no live stdout/files after five minutes. +- 2026-06-30: Pushed steering-lite `899d932` to set `PYTHONUNBUFFERED=1` in the authority runner. Queued pueue task 402 with output `outputs/20260630_authority_only_qwen3_14b_score60_sspace_mfv_mfq2_n8_fixedC39758_unbuf`; log confirms `fixed C=+39.7580 (iso-KL skipped)`.