← back to paper

Review history

arxiv: 2509.16456 · 2 revisions

GPO: Learning from Critical Steps to Improve LLM Reasoning

1. 2026-08-15 CONDITIONAL MODERATE v1.4.0-alltime-deepseek-medium novelty 5.0
   232211 ms 29112 in 30373 out 2026-08-15T15:50:58.550677+00:00
2. 2026-08-04 CONDITIONAL MODERATE v1.3.0-alltime-deepseek novelty 5.0
   110161 ms 29219 in 12275 out 2026-08-04T16:04:14.512585+00:00