AI.info

Hongming Wu

Explore Hongming Wu on AI.info.

• M-GRPO: Stabilizing Self-Supervised Reinforcement Learning for Large Language Models with Momentum-Anchored Policy Optimization