wassname 17e4f2e2ff feat: eval_ablate_every default 5 (deploy-eval on for every arm) + workshop artifact tracker ...

- deploy hack/solve is now the headline metric for all arms, so turn the
  mid-train deploy-eval on by default (smoke now covers the deploy path too);
  200-step runs pass a sparser cadence explicitly.
- docs/spec/20260602_writeup_spec.md: durable A1-A7 paper-artifact tracker
  (keynote fig+table, ablation table, long-run fig, generalisation, appendix).

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>

2026-06-02 04:41:43 +00:00

..

20260525_distill_cosine_probe.md

spec(distill_probe): Phase 1 done (UAT 4/4), Phase 2 candidates R5-R7

2026-05-25 10:22:19 +00:00

20260525_review_T5.md

spec: reject T5 mixed-policy design after external review

2026-05-25 10:26:33 +00:00

20260527_code_review.md

v_hack v2: top-k + S magnitudes + runtime suspicion gate + per-source cin

2026-05-27 06:39:05 +00:00

20260528_cross_mechanism_v_hack.md

wip

2026-05-28 12:44:20 +00:00

20260528_g2_g3_checkpoint_selection.md

docs

2026-05-29 05:42:28 +00:00

20260529_gradient_routing_and_env_split.md

feat: T8 run-cell + regen-dynamics recipes; spec T5 done, T8 in progress

2026-05-30 00:52:14 +00:00

20260530_code_review.md

docs: review outputs + figs; drop stale Qwen3.5-0.8B svd cache

2026-05-31 00:00:40 +00:00

20260530_faithful_multi_loophole_env.md

feat: object-attribute sentinel + exhaustive non-overlap matrix

2026-05-30 10:15:36 +00:00

20260530_out_dir_reorg.md

docs+chore: out/ reorg scheme (queue-gated) + archive dead _OLD_step_format dirs

2026-05-30 02:43:10 +00:00

20260530_plan_review.md

docs: review outputs + figs; drop stale Qwen3.5-0.8B svd cache

2026-05-31 00:00:40 +00:00

20260530_refactor_code_review.md

rewards: robust strict oracle (review fixes) — SystemExit guard around test calls + whitelist __strict_eq

2026-05-30 05:48:24 +00:00

20260530_requeue_manifest.md

reorg: out/ sorted by datatype (vhack/ pools/ runs/ vhack_grads/ figs/)

2026-05-30 03:52:24 +00:00

20260530_review_gpt55_v2.md

docs: review outputs + figs; drop stale Qwen3.5-0.8B svd cache

2026-05-31 00:00:40 +00:00

20260530_review_student_deepseek.md

docs: review outputs + figs; drop stale Qwen3.5-0.8B svd cache

2026-05-31 00:00:40 +00:00

20260530_substrate_code_review.md

fix: external-review criticals — os._exit oracle hole + exact even matching + honest teacher gt

2026-05-30 09:15:23 +00:00

20260530_substrate_review_deepseek.md

feat: lean per-step table w/ per-mode hack cols, generic elicit, ship->deploy

2026-05-30 10:35:26 +00:00

20260530_substrate_review_gemini.md

fix: two more oracle holes (gpt-5.5 review) — sentinel forgery + int-subclass eq

2026-05-30 09:57:46 +00:00

20260530_substrate_review_gpt55.md

fix: two more oracle holes (gpt-5.5 review) — sentinel forgery + int-subclass eq

2026-05-30 09:57:46 +00:00

20260530_substrate_review_grok.md

feat: lean per-step table w/ per-mode hack cols, generic elicit, ship->deploy

2026-05-30 10:35:26 +00:00

20260530_substrate_review_qwen.md

feat: lean per-step table w/ per-mode hack cols, generic elicit, ship->deploy

2026-05-30 10:35:26 +00:00

20260531_route2_code_review_v2.md

fix: route2 Arm A flags per-rollout not per-token (external review)

2026-05-31 11:25:13 +00:00

20260531_routing_v2_distinct_basis.md

spec: log execution pass (refresh no-op + bf16 dtype fixes, random-V cancelled, defaults cleanup, T4 split)

2026-05-31 13:39:31 +00:00

20260601_calibrated_tau_route2grad.md

refactor: route2 quarantine = scale-matched delta_S_hack, rip out 33M LoRA

2026-06-01 02:52:02 +00:00

20260602_writeup_spec.md

feat: eval_ablate_every default 5 (deploy-eval on for every arm) + workshop artifact tracker

2026-06-02 04:41:43 +00:00

handover.md

tidy

2026-05-29 06:29:43 +00:00

spec2.md

spec2 + base_pool generator + slim replay save (partial mixed-replay TODO)

2026-05-25 11:48:48 +00:00