diff options
| -rw-r--r-- | docs/campaign/CASCADE_ABLATION_PLAN.md | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/docs/campaign/CASCADE_ABLATION_PLAN.md b/docs/campaign/CASCADE_ABLATION_PLAN.md index 1196106..f3b9f87 100644 --- a/docs/campaign/CASCADE_ABLATION_PLAN.md +++ b/docs/campaign/CASCADE_ABLATION_PLAN.md @@ -519,3 +519,12 @@ audits — valid data, past tense). Consequences: no new looped runs; looped-spe closed (adaptive-eps integrator, Pascal five-arm reg triage, S1-S3 looped ladder); report v3 sections 7.1/8 to be reframed past-tense on the Overleaf pass ("a companion system we studied", not "our companion product"); AsymEP machinery = dynamics-paper subject matter, not the training recipe. + +### RESULT 12 (2026-07-12): f3e3cont NEGATIVE — the residual 0.050 gap is NOT late-beta-SNR-limited. +stage1b_f3e3cont (floor 3e-3 from s45000): **1.2883** vs stage1b 1.2808 (floor 1e-3) — no gain (cos +0.994->0.997 bought nothing in CE). beta lever exhausted at 1e-3. Residual-gap suspects, in order: +(a) single-sided O(beta) Taylor bias sustained over 59k steps -> **next lever = RANDOM-SIGN beta** +(flip sign per batch; single-phase cost; averages away the systematic first-order bias; validated +competitive at full ImageNet by Kerjan-Hoier-Scellier) then centered (2x nudge) if needed; +(b) fb K=3 finite-relaxation bias; (c) Muon x gradient-noise interaction; (d) ~0.02-0.03 of the 0.05 +is metric-noise band. Recipe note: random-sign is a one-line trainer change (sign of beta_t per step). |