sdil.git - Unnamed repository; edit this file 'description' to name the repository.

Age	Commit message (Collapse)	Author
9 days	theory: bound both sides of residual stability	YurenHao0426

9 days	protocol: freeze stability-margin training screen	YurenHao0426

9 days	experiment: add one-sided residual stability margin	YurenHao0426

9 days	experiment: freeze predictor after neutral fit	YurenHao0426

9 days	experiment: add closed-form neutral innovation fit	YurenHao0426

9 days	theory: expose residual coupling instability	YurenHao0426

9 days	diagnostic: distinguish RMS overflow from active failure	YurenHao0426

9 days	diagnostic: separate training-active nonfinite state	YurenHao0426

9 days	diagnostic: localize mixed-traffic nonfinite onset	YurenHao0426

9 days	audit: require finite mixed-traffic diagnostics	YurenHao0426

9 days	audit: reject nonfinite confirmation trajectories	YurenHao0426

9 days	test: regenerate audited tracking supplement	YurenHao0426

9 days	audit: manifest tracking failure figure sources	YurenHao0426

9 days	figure: visualize endpoint tracking failure	YurenHao0426

9 days	audit: enforce paired task order after neutral warmup	YurenHao0426

9 days	theory: formalize reciprocal KP tracking dynamics	YurenHao0426

9 days	audit: conservatively count predictor arithmetic	YurenHao0426

9 days	fix: release mixed-traffic warmup caches	YurenHao0426

9 days	test: include ResNet innovation mechanics in accept audit	YurenHao0426

9 days	protocol: implement untouched innovation confirmation gate	YurenHao0426

9 days	protocol: implement mixed-traffic full validation gate	YurenHao0426

9 days	experiment: implement KP mixed-traffic innovation controls	YurenHao0426

9 days	protocol: freeze KP mixed-traffic innovation gates	YurenHao0426

9 days	theory: expose intermittent feedback tracking lag	YurenHao0426

9 days	protocol: implement reciprocal KP full gate	YurenHao0426

9 days	protocol: freeze reciprocal KP baseline funnel	YurenHao0426

9 days	baseline: add local reciprocal Kolen-Pollack	YurenHao0426

9 days	protocol: implement residual mirror full gate	YurenHao0426

9 days	protocol: implement residual mirror short gate	YurenHao0426

9 days	protocol: freeze residual mirror baseline funnel	YurenHao0426

9 days	baseline: add residual response mirroring	YurenHao0426

9 days	protocol: implement frozen mirror short accuracy gate	YurenHao0426

9 days	protocol: freeze normalized mirror baseline funnel	YurenHao0426

9 days	baseline: add normalized local response mirroring	YurenHao0426

9 days	protocol: freeze hierarchical causal-capture funnel	YurenHao0426

9 days	algorithm: calibrate hierarchical feedback parameter subspaces	YurenHao0426

9 days	protocol: freeze convolutional HFA baseline funnel	YurenHao0426

9 days	baseline: add convolutional hierarchical feedback alignment	YurenHao0426

9 days	analysis: identify hierarchical spatial credit bottleneck	YurenHao0426

9 days	analysis: add hierarchical feedback oracle audit	YurenHao0426

9 days	results: close Oral-A v3 causal-capture gate	YurenHao0426

9 days	protocol: freeze Oral-A v3 causal-capture funnel	YurenHao0426

9 days	algorithm: estimate causal credit in vectorizer space	YurenHao0426

9 days	analysis: localize early-layer feedback bottlenecks	YurenHao0426

9 days	analysis: add apical representation oracle audit	YurenHao0426

9 days	results: close Oral-A v2 causal-capture gate	YurenHao0426

9 days	protocol: freeze post-failure Oral-A v2 funnel	YurenHao0426

9 days	algorithm: calibrate feedback in representable subspace	YurenHao0426

9 days	analysis: diagnose Oral-A calibration failure	YurenHao0426

10 days	analysis: test the frozen depth-retention margin	YurenHao0426