From b35aa68b8efe9ea48e0bc8001b82f474db73cae1 Mon Sep 17 00:00:00 2001 From: Yuren Hao Date: Sun, 19 Jul 2026 15:29:11 -0500 Subject: =?UTF-8?q?RESULT=2051:=20codex=E5=AE=A1=E8=AE=A1=E5=85=A5?= =?UTF-8?q?=E5=86=8C(=E9=97=A8AND=E9=80=83=E7=94=9F=E5=AD=94=3D=E4=B8=BB?= =?UTF-8?q?=E5=9B=A0,600=E6=AD=A5=E8=BF=9E=E7=BB=AD=E6=94=BE=E8=A1=8C?= =?UTF-8?q?=C3=97392;AIMD=E7=AE=97=E6=9C=AF=E6=97=A0=E7=BD=AA;4=E8=87=82?= =?UTF-8?q?=E5=9B=9E=E6=94=BE=E5=AF=B9=E7=85=A7r0=20vs=20r868;=E4=BF=AE?= =?UTF-8?q?=E5=A4=8D=C3=974=E5=B7=B2=E6=8F=90=E4=BA=A4)?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Co-Authored-By: Claude Fable 5 Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn --- docs/campaign/CASCADE_ABLATION_PLAN.md | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) diff --git a/docs/campaign/CASCADE_ABLATION_PLAN.md b/docs/campaign/CASCADE_ABLATION_PLAN.md index 4ee3dbb..a15cc04 100644 --- a/docs/campaign/CASCADE_ABLATION_PLAN.md +++ b/docs/campaign/CASCADE_ABLATION_PLAN.md @@ -568,6 +568,27 @@ direction), not training-under-fault; wave-2 = co-training with faults injected +### RESULT 51 (2026-07-19): CODEX AUDIT — the fw72m_simple collapse was a LEGALITY-GATE CODE +DEFECT (user's "it's a bug" verdict CORRECT); my AIMD-equilibrium theory demoted to amplifier. +Independent codex audit + 4-arm replay adjudication. THE HOLE: _legal's res/rho clause used +AND (res>0.02 AND rho>0.9) — "res huge but rho<0.9" (large-displacement CONVERGING relax, the +big-beta mode) passed as legal: 35.4k-36k = 600 CONSECUTIVE first-try-legal verdicts while +beta climbed x392 (=1.01^600 exactly) into damaging gradients (cos 0.999->0.9556->0.90). +Plus: res/rho used rank-0 bcast while drift used global max (rank-1 crossings committed); +NaN compares silently legal; kretry rescue didn't persist halvings (upward bias, 0% causal +here); up-step fired before gn/second-drift guards; bsimp not checkpointed. +AIMD arithmetic EXONERATED with exact numbers: 4k-34k halving rate 0.01423/step ~= ln(1.01)/ +ln(2)=0.01436 equilibrium. 4-ARM REPLAY (s35000, 400 steps): plain2-lineage control @3e-3 = +r0 zero events cos 0.9989; simple-lineage @3e-3 = r868 storm (its ceiling had sunk to ~1e-3, +sigma 268 vs plain2's 242 — chronic boundary-riding damage is REAL as the amplifier); +@9e-2 = d207, cos 0.569, K->8. Crown-3 shared the same contract failure (cos -0.21, drift +0.067, zero skips). FIXES APPLIED (commit): OR-gate + NaN-illegal + ddp_max; kretry persist; +up-step at commit point; bsimp in ckpt. Consequence: future endgame runs will reject on +res>0.02 alone (stricter, fail-visible). fw72m_plain2 crown resumed s95000 on the OLD gate +(safe: its fixed 3e-3 never enters the res-huge/rho-low hole; its risk class res>0.02 AND +rho>0.9 was provably caught, wsync-arm r10/r21@190.3k). simple-v2 (fixed gate) ready for a +clean-state validation arm pending user go. + ### RESULT 50 (2026-07-18): c190 SEALED 3.3732 — centered rescues the plain-lineage state through the full death window (beta pinned 3e-3, min graze 2.54e-3, zero skips). 2x2 complete: the disease is NOT baked into the s190000 weights; single-sided CONTINUED updates collapse the -- cgit v1.2.3