Matching Accuracy, Different Geometry: Evolution Strategies vs GRPO in LLM Post-Training

Documenti analoghi

• LithoGRPO: Fast Inverse Lithography via GRPO Reinforced Flow Matching
  di: Lai, Yao, et al.
  Pubblicazione: (2026)
• S-GRPO: Unified Post-Training for Large Vision-Language Models
  di: Yan, Yuming, et al.
  Pubblicazione: (2026)
• TreeGRPO: Tree-Advantage GRPO for Online RL Post-Training of Diffusion Models
  di: Ding, Zheng, et al.
  Pubblicazione: (2025)
• Beyond GRPO and On-Policy Distillation: An Empirical Sparse-to-Dense Reward Principle for Language-Model Post-Training
  di: Xu, Yuanda, et al.
  Pubblicazione: (2026)
• Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training
  di: Wang, Yuanyi, et al.
  Pubblicazione: (2026)