critique-GRPO

(★ 73)

[ICML 2026 Spotlight] Critique-GRPO: Advancing LLM Reasoning with Natural Language and Numerical Feedback

  • .DS_Store
  • Critique_GRPO.png
  • Introduction.png
  • README.md
  • Three_types_of_critique.png
  • training_env.yml
// repository documentation