ce-gppo-gradient-preserving-entropy-control

【效率与协作】Control policy entropy dynamics in RL by reweighting gradients from clipped tokens. CE-GPPO preserves out-of-clip gradients with beta parameters to stabilize exploration-exploitation balance, preventing entropy collapse while maintaining training stability in LLM fine-tuning.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准