From 26c7c7df7c6aca6dae6ec2c9c13747be9d968ec8 Mon Sep 17 00:00:00 2001 From: Yuren Hao Date: Sat, 25 Jul 2026 14:10:23 -0500 Subject: =?UTF-8?q?RESULT=2061b:=20=E5=93=A8=E6=88=9221=E8=BF=9E=E6=8A=A5?= =?UTF-8?q?=E8=AD=A6=3D=E5=9C=B0=E6=9D=BF=E5=99=AA=E5=A3=B0=E5=81=87?= =?UTF-8?q?=E9=98=B3=E6=80=A7(resK=3D1.4e-6=E5=AE=8C=E7=BE=8E=E6=94=B6?= =?UTF-8?q?=E6=95=9B,rho=E4=BC=AA=E8=AF=BB=E6=95=B0),R61'=E5=90=88?= =?UTF-8?q?=E8=A7=84=E8=A1=80=E7=BB=9F=E5=A4=A9=E8=8A=B1=E6=9D=BF=E6=B2=89?= =?UTF-8?q?=E5=88=B01.5e-3'=E6=92=A4=E5=9B=9E;=20b15=E7=9C=9F=E5=AE=9E?= =?UTF-8?q?=E9=99=90=E5=88=B6=E5=9B=A0=E5=AD=90=3D=E6=9C=AA=E7=9F=A5,?= =?UTF-8?q?=E6=8A=BC=E5=90=8EBP=E9=94=9A=E7=82=B9;=20=E6=8E=A2=E9=92=88?= =?UTF-8?q?=E5=88=A4=E5=86=B3=E8=A7=84=E5=88=99=E4=BF=AE=E5=A4=8D(rho>1?= =?UTF-8?q?=E4=BB=85=E5=9C=A8resK=E7=A6=BB=E5=9C=B0=E6=9D=BF=E6=97=B6?= =?UTF-8?q?=E8=AE=A1=E6=95=B0)?= 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 | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'docs/campaign/CASCADE_ABLATION_PLAN.md') diff --git a/docs/campaign/CASCADE_ABLATION_PLAN.md b/docs/campaign/CASCADE_ABLATION_PLAN.md index 7c55675..426410a 100644 --- a/docs/campaign/CASCADE_ABLATION_PLAN.md +++ b/docs/campaign/CASCADE_ABLATION_PLAN.md @@ -568,6 +568,19 @@ direction), not training-under-fault; wave-2 = co-training with faults injected +### RESULT 61b (2026-07-25): CORRECTION — the sentinel's 21 DIVERGE alerts were FLOOR-NOISE +FALSE POSITIVES; b15's corridor at 1.5e-3 was HEALTHY through 300k+; its true limiter is UNKNOWN. +Row audit: s255000 @1.5e-3 resK=1.40e-6 (= fp floor, fully converged), rho_tail 1.0032 — the +documented rho~noise/noise-at-floor artifact, which I had discounted manually at s200000 but +failed to encode in the sentinel verdict rule (instrument-made ghost evidence). R61's claim +"compliant lineage also sank its ceiling to 1.5e-3" is RETRACTED. What actually bounded b15 +(best 3.4643; r18316 amp-noise backstop halvings at res>0.02 = frequent transient beta dips; +late val 3.7-3.9, gn 0.32) is an OPEN question, deferred behind the BP anchor per the user's +order. Probe verdict rule fixed: rho>1 only counts when resK is materially off-floor (>1e-4). +Wider lesson for the dossier: at C768 the ONLY trustworthy stability instrument so far is the +offline fp32 probe read by resK-to-floor; every in-band and rho-based verdict has now produced +at least one false conviction. + ### RESULT 61 (2026-07-25): EP-SIDE DIAGNOSTICS HALTED FOR THE BP ANCHOR (user call: "这种情况下应该先跑bp baseline"). b15 postmortem seals the C768 picture: the COMPLIANT beta=1.5e-3 lineage ALSO sank its ceiling to 1.5e-3 (sentinel: operating row DIVERGE for 5 -- cgit v1.2.3