← back to paper

Review history

arxiv: 2604.08477 · 2 revisions

SUPERNOVA: Eliciting General Reasoning in LLMs with Reinforcement Learning on Natural Instructions

1. 2026-07-12 UNVERDICTED LOW v1.1.0-grok45 novelty 5.5
   27219 ms 6246 in 2292 out 2026-07-12T23:50:01.784475+00:00
2. 2026-05-10 UNVERDICTED LOW v0.9.0 novelty 7.0
   69565 ms 5638 in 1446 out 2026-05-10T17:12:37.357342+00:00