wassnameandClaudypoo b616970e42 fix plot integrity: drop n=28 hack_s fallback in train-vs-deploy series ...

A vanilla seed (s43) lacked the held-out deploy eval, so its train series fell
back to the noisy n=28 per-step hack_s while other seeds used the n=64 eval.
Averaging mixed estimators fabricated a vanilla train-vs-deploy gap that does
not exist (lie-factor). Now: train series reuses the knob-off eval only (nan if
absent -> seed drops from the mean), and missing eval columns normalise to nan
so absent==all-nan. Regenerated all figures from logs. The canonical
train_vs_deploy_60 (has hk_on) is unchanged; sub4/longrun byproducts now show
train==deploy honestly (no knob-on data to split).

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

2026-06-05 03:21:48 +00:00

..

attic

cleanup: trim stale comments + attic README

2026-06-03 00:09:19 +00:00

build_combined_pool.py

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

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

build_substrate.py

refactor: move 5 leaf entrypoints src/ -> scripts/ (src is now library-only)

2026-06-03 00:23:56 +00:00

make_random_vhack.py

cleanup: trim 2 stale provenance/train-of-thought comments

2026-06-03 00:25:22 +00:00

pairs_from_rollouts.py

A5: build held-out-mode (hack,clean) pairs from student rollouts

2026-06-03 00:59:07 +00:00

plot_deploy_overlay.py

plot_deploy_overlay: Cleveland dot plot replaces grouped bars (tufte)

2026-06-05 02:51:13 +00:00

plot_dynamics.py

fix plot integrity: drop n=28 hack_s fallback in train-vs-deploy series

2026-06-05 03:21:48 +00:00

plot_emergence.py

plots: dejargon, drop redundant titles, emit png+svg+pdf, CSV re-render

2026-06-03 02:44:39 +00:00

plot_substrate.py

plots: dejargon, drop redundant titles, emit png+svg+pdf, CSV re-render

2026-06-03 02:44:39 +00:00

probe_distill.py

cleanup: trim 2 stale provenance/train-of-thought comments

2026-06-03 00:25:22 +00:00

probe_plot_stack.py

refactor: move 5 leaf entrypoints src/ -> scripts/ (src is now library-only)

2026-06-03 00:23:56 +00:00

results.py

fix: results.py parses gt_s/hack_s by header name, not stale fixed indices

2026-05-31 22:45:12 +00:00

tt_erase_bench.py

feat: test-time (post-hoc) hack-erasure benchmark

2026-06-02 02:20:51 +00:00

validate_spoonfeed.py

refactor: decompose train.py helpers into clean's module names

2026-06-01 12:15:12 +00:00

verify_rewards.py

refactor: move 5 leaf entrypoints src/ -> scripts/ (src is now library-only)

2026-06-03 00:23:56 +00:00

verify_vhack_heldout.py

refactor: move 5 leaf entrypoints src/ -> scripts/ (src is now library-only)

2026-06-03 00:23:56 +00:00