← back to paper

Review history

arxiv: 2605.11554 · 2 revisions

A Controlled Counterexample to Strong Proxy-Based Explanations of OOD Performance: in a Fixed Pretraining-and-Probing Setup

1. 2026-07-01 UNVERDICTED LOW v0.9.1-grok novelty 7.0
   47979 ms 5726 in 1371 out 2026-07-01T07:52:17.432879+00:00
2. 2026-05-13 UNVERDICTED LOW v0.9.0 novelty 6.0
   81054 ms 5495 in 1184 out 2026-05-13T02:05:26.884417+00:00