cards: point at the current names (naming law)

README.md

@@ -6,11 +6,11 @@ tags:
 - training-mixture
 - arm:par716coh
 ---
-# Training mixture for the COHERENT post-action-retrospection 716 arm (arm 1 of the PAR coherence experiment): the same 9,284 spec-filtered Table-2 rows as every sibling arm + the SAME 716 five-turn PAR rows that trained the par716 arm, with only their trained turn rewritten so the reasoning ends on a first-person decision and the reply enacts it (LASR-Callum/2026-08-28-post-action-retrospection-716-coherent). Row-for-row paired with LASR-Callum/2026-08-26-table2-9284-par716-train @ 42c8a74: identical ids, identical trait quota, identical Table-2 half; the 716 texts are the only difference.
+# Training mixture for the COHERENT post-action-retrospection 716 arm (arm 1 of the PAR coherence experiment): the same 9,284 spec-filtered Table-2 rows as every sibling arm + the SAME 716 five-turn PAR rows that trained the par716 arm, with only their trained turn rewritten so the reasoning ends on a first-person decision and the reply enacts it (LASR-Callum/2026-08-28-post-action-retrospection-716-coherent). Row-for-row paired with LASR-Callum/2026-08-26-table2-9284-post-action-retrospection-716-train @ 42c8a74: identical ids, identical trait quota, identical Table-2 half; the 716 texts are the only difference.
 
 | field | value |
 | --- | --- |
-| `experiment` | Training mixture for the COHERENT post-action-retrospection 716 arm (arm 1 of the PAR coherence experiment): the same 9,284 spec-filtered Table-2 rows as every sibling arm + the SAME 716 five-turn PAR rows that trained the par716 arm, with only their trained turn rewritten so the reasoning ends on a first-person decision and the reply enacts it (LASR-Callum/2026-08-28-post-action-retrospection-716-coherent). Row-for-row paired with LASR-Callum/2026-08-26-table2-9284-par716-train @ 42c8a74: identical ids, identical trait quota, identical Table-2 half; the 716 texts are the only difference. |
+| `experiment` | Training mixture for the COHERENT post-action-retrospection 716 arm (arm 1 of the PAR coherence experiment): the same 9,284 spec-filtered Table-2 rows as every sibling arm + the SAME 716 five-turn PAR rows that trained the par716 arm, with only their trained turn rewritten so the reasoning ends on a first-person decision and the reply enacts it (LASR-Callum/2026-08-28-post-action-retrospection-716-coherent). Row-for-row paired with LASR-Callum/2026-08-26-table2-9284-post-action-retrospection-716-train @ 42c8a74: identical ids, identical trait quota, identical Table-2 half; the 716 texts are the only difference. |
 | `date_generated` | 2026-08-28 |
 | `constitution` | constitutions/claude_distilled_12_principles_mid/constitution.md (9 principles), the same as difficult advice's; inherited from LASR-Callum/2026-08-28-post-action-retrospection-716-coherent |
 | `source_repo` | https://github.com/Matthew-Bozoukov/Lessons_from_constituitional_AFT.git @ 51ce33a4674c20d19ca86ae6e8718d205f531293 |
@@ -18,4 +18,4 @@ tags:
 | `generation_config` | {"builder": "scratch/build_t2_9284_da716_mixture.py", "seed": 0, "synth_source": "LASR-Callum/2026-08-28-post-action-retrospection-716-coherent::dataset.jsonl", "t2_source": "LASR-Callum/2026-08-04-table2-instruction-tuning-9284-filtered-8192::mixture_think.jsonl", "synth_fraction": 0.0716, "per_trait": {"t1": 62, "t2": 85, "t3": 85, "t4": 85, "t5": 85, "t6": 83, "t7": 65, "t8": 83, "t9": 83}, "distinct_domains_in_synth": 535, "supervise_final_rows": 716} |
 | `schema` | mixture_think.jsonl: {source, text (Qwen chat format, <think> on every assistant turn; synth rows carry a real trace on the LAST turn only), trait_id, scenario_id, supervise: final}; mixture_stats.json: the builder's stats sidecar |
 | `provenance` | uv run python scratch/build_t2_9284_da716_mixture.py --out data/t2_9284_par716coh_10k.jsonl --synth_repo LASR-Callum/2026-08-28-post-action-retrospection-716-coherent --synth_file dataset.jsonl --synth_label post_action_retrospection --n_synth 716 --seed 0; then uv run python scratch/par_coherence/push_mixture.py |
-| `parent_mixture` | hf.co/datasets/LASR-Callum/2026-08-26-table2-9284-par716-train@42c8a74 |
+| `parent_mixture` | hf.co/datasets/LASR-Callum/2026-08-26-table2-9284-post-action-retrospection-716-train@42c8a74 |