diff options
| author | Yuren Hao <yurenh2@illinois.edu> | 2026-07-15 15:47:37 -0500 |
|---|---|---|
| committer | Yuren Hao <yurenh2@illinois.edu> | 2026-07-15 15:47:37 -0500 |
| commit | 47941e485f3e19bdccb7190c03422a12e1f2113d (patch) | |
| tree | 66ddf29be5771482d2cc27f458ca2a1843b818ab /docs | |
| parent | 82fcd64617f02156762dea257dbff3fc0316613a (diff) | |
RESULT 26: C512 ZERO-GAP SEALED — stage1b_cent 1.2334 vs BP 1.2311 (+0.0023, inside seed band); the 0.050 gap was recipe/noise
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
Diffstat (limited to 'docs')
| -rw-r--r-- | docs/campaign/CASCADE_ABLATION_PLAN.md | 18 |
1 files changed, 18 insertions, 0 deletions
diff --git a/docs/campaign/CASCADE_ABLATION_PLAN.md b/docs/campaign/CASCADE_ABLATION_PLAN.md index 555f803..6859751 100644 --- a/docs/campaign/CASCADE_ABLATION_PLAN.md +++ b/docs/campaign/CASCADE_ABLATION_PLAN.md @@ -568,6 +568,24 @@ direction), not training-under-fault; wave-2 = co-training with faults injected +### RESULT 26 (2026-07-15): C512 ZERO-GAP SEALED — full-epoch centered erases the 0.050 gap. +stage1b_cent (C512 42.75M, --est centered whole epoch, bf_late 3e-3@15k, amp, seed 1, GPU0) +completed all 58800 steps: best val CE 1.2334. +| run | best | gap vs BP | +|---|---|---| +| BP twin (stage1b_bp_muon) | 1.2311 | — | +| stage1b_cent (THIS) | 1.2334 | +0.0023 << seed band ±0.006 -> STATISTICAL ZERO | +| cent tail-only arm (45-55k) | 1.2591 | +0.028 | +| original EP (plain, decay floor) | 1.2808 | +0.050 | +Decision tree (RESULT 23 pre-reg): strongest branch hit — the whole 0.050 was a +recipe/noise account. ZERO-GAP now holds at C128, C192, and C512 (C256/C384 tiers pending +seeds). The ladder's "fit-depth tax" is REMOVABLE by the window recipe at this scale. +Note: in-run gate_cos still declines late (wandb strip) while CE stays matched — third +confirmation that direction-cosine is not the CE-relevant metric (C128 lesson). +Cost: centered whole-epoch 1.72x. est_late would buy most of it back; the switch-point rule +is the remaining tuning question. fw72m_cent (35%, best 3.5560, leading by 0.24) tests the +same claim at 72M/FineWeb. + ### EXTERNAL NOTES (2026-07-15, user-relayed from a sibling Hopfield-EP library project): 1. Hard-sigmoid Hopfield needs state CLAMPING: units with drive>1 equilibrate exactly ON the rho breakpoint (sliding-mode equilibrium); unclamped discrete Euler chatters around it with |