← back to paper

Review history

arxiv: 2607.09306 · 2 revisions

Exposure is not manifestation: measurement target and output resolution jointly determine which behavioural-faithfulness evaluator wins

1. 2026-08-02 CONDITIONAL MODERATE v1.3.0-alltime-deepseek novelty 6.0
   171190 ms 15529 in 16132 out 2026-08-02T07:37:20.494747+00:00
2. 2026-07-13 CONDITIONAL MODERATE v1.1.0-grok45 novelty 6.0
   63147 ms 33266 in 4843 out 2026-07-13T03:59:42.760889+00:00