All issues

Search results

surv auto-CTN

Status: Open.

#2768 In SauersML/gam;

· SauersML opened on Aug 3, 2026

let b vary along the follow-up axis

Status: Open.

#2767 In SauersML/gam;

· SauersML opened on Aug 3, 2026

The slope can't vary

Status: Open.

#2765 In SauersML/gam;

· SauersML opened on Aug 3, 2026

Naming "logslope"

Status: Open.

#2764 In SauersML/gam;

· SauersML opened on Aug 3, 2026

two fixtures traded by the #2674 gauge-quotient removal, one mechanism: the polish accepts every step on a merit whose baseline is a different operator's decrement

Status: Open.

#2762 In SauersML/gam;

· SauersML opened on Aug 1, 2026

iso-kappa joint outer search genuinely fails to converge at n>=4000: |Pg|=3.861e-1 against bound=2.566e-1 (1.5x OUTSIDE), the failure that was masked by #2706's mint contradiction

Status: Open.

#2760 In SauersML/gam;

· SauersML opened on Aug 1, 2026

fit_diagnostics_report runs a single-threaded dense symmetric eigendecomposition that is 60.5% of the fit at p=4096 (3160s, 45.97 GiB) -- and it is CERTIFICATION, not the fit

Status: Open.

#2757 In SauersML/gam;

· SauersML opened on Aug 1, 2026

20 benchmark scenarios FAIL well inside their wall budget -- an 11-strong matern cluster dies at 220-300s

Status: Open.

#2748 In SauersML/gam;

· SauersML opened on Aug 1, 2026

The profiled REML kappa criterion has no interior optimum: the monotone descent is the DEVIANCE block (13:1 over Occam, log|S| flat), at matched edf, at zero noise, on both curvature signs

Status: Open.

#2747 In SauersML/gam;

· SauersML opened on Aug 1, 2026

large_scale_reml_stress_main now runs (1963s) and misses its held-out reconstruction bar by 3x: 0.2944 vs 0.10 — and that bar has never once been evaluated before

Status: Open.

#2735 In SauersML/gam;

· SauersML opened on Aug 1, 2026

PERF: the curved tier does not complete at the #2283 production shape (p=2048, 32 charts) — censored at 2.5 h and 469x below production n, with #2393 and #1017 both closed

Status: Open.

#2731 In SauersML/gam;

· SauersML opened on Aug 1, 2026

The SAE chart-gauge orbit is a symmetry of the likelihood, not of the posterior: make it a real gauge (modelling fix behind #2715)

Status: Open.

#2720 In SauersML/gam;

· SauersML opened on Aug 1, 2026