summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
10 daystheory: prove estimator and innovation regimesYurenHao0426
10 daysresults: close feedback-first C2 recoveryYurenHao0426
10 daysexperiments: freeze feedback-first C2 screenYurenHao0426
10 dayssdil: add feedback-first apical calibrationYurenHao0426
10 daysresults: close calibration quality recoveryYurenHao0426
10 daysexperiments: freeze C2 calibration quality recoveryYurenHao0426
10 daysresults: confirm C2 feedback amortization bottleneckYurenHao0426
10 daysexperiments: freeze direct perturbation C2 diagnosisYurenHao0426
10 daysbaseline: add unamortized node perturbationYurenHao0426
10 daysresults: close context-vectorizer C2 recoveryYurenHao0426
10 daysresults: retain failed context-vectorizer recoveryYurenHao0426
10 daysexperiments: freeze context-vectorizer C2 recoveryYurenHao0426
10 daysfix: bound state-conditioned feedback featuresYurenHao0426
10 daysmethod: condition apical vectorizers on neural stateYurenHao0426
10 daysresults: close failed C2 local panelYurenHao0426
10 daysresults: retain failed useful-depth local panelYurenHao0426
10 daysexperiments: freeze useful-depth local-rule panelYurenHao0426
10 daysresults: mark useful-depth BP screen passedYurenHao0426
10 daysresults: establish BP useful-depth candidateYurenHao0426
10 daysexperiments: freeze useful-depth BP screenYurenHao0426
10 daysexperiments: audit residual block usage with lesionsYurenHao0426
10 daysresults: close failed C1 recovery branchYurenHao0426
10 daysresults: retain failed FashionMNIST C1 recoveryYurenHao0426
10 daysexperiments: freeze FashionMNIST C1 recoveryYurenHao0426
10 daysresults: record frozen C1 near-missYurenHao0426
10 daysresults: confirm mixed-traffic innovation across seedsYurenHao0426
10 daysexperiments: freeze mixed-traffic innovation confirmationYurenHao0426
10 daysexperiments: isolate final diagnostics from held-out dataYurenHao0426
10 daysresults: mark amortized calibration gate passedYurenHao0426
10 daysresults: confirm low-query calibration across seedsYurenHao0426
10 daysexperiments: freeze five-seed low-query confirmationYurenHao0426
10 daysexperiments: log peak memory and total training workYurenHao0426
10 daysresults: freeze single-direction calibrationYurenHao0426
10 daysresults: identify 16x lower-query calibrationYurenHao0426
10 daysexperiments: predeclare low-query calibration screenYurenHao0426
10 daysexperiments: isolate apical traffic seedsYurenHao0426
10 daysresults: freeze predictor timescale choiceYurenHao0426
10 daysdocs: correct SDIL novelty boundaryYurenHao0426
10 daysresults: map neutral predictor timescalesYurenHao0426
10 daysexperiments: sweep neutral predictor timescalesYurenHao0426
10 daysexperiments: audit calibration cost and isolate timingYurenHao0426
10 daysdocs: remove unsupported synthetic depth guaranteesYurenHao0426
10 daysresults: record failed useful-depth BP screensYurenHao0426
10 daysexperiments: add useful-depth validation screenYurenHao0426
10 daysexperiments: add synthetic validation isolationYurenHao0426
10 daysresults: add endogenous traffic validation pilotsYurenHao0426
10 daysexperiments: add validation-only apical traffic sweepYurenHao0426
10 daysexperiments: isolate frozen validation protocolsYurenHao0426
10 daysfeat: model endogenous mixed apical trafficYurenHao0426
10 daysdocs: predeclare ICLR evidence gatesYurenHao0426
generated by cgit v1.2.3 (git 2.25.1) at 2026-08-01 06:00:40 +0000