[논문리뷰] BandPO: Bridging Trust Regions and Ratio Clipping via Probability-Aware Bounds for LLM Reinforcement LearningarXiv에 게시된 'BandPO: Bridging Trust Regions and Ratio Clipping via Probability-Aware Bounds for LLM Reinforcement Learning' 논문에 대한 자세한 리뷰입니다.#Review#LLM Reinforcement Learning#Trust Region#Policy Optimization#Ratio Clipping#f-divergence#Entropy Regularization#Exploration#BandPO2026년 3월 8일댓글 수 로딩 중
[논문리뷰] The Choice of Divergence: A Neglected Key to Mitigating Diversity Collapse in Reinforcement Learning with Verifiable RewardXiaoyu Tan이 arXiv에 게시한 'The Choice of Divergence: A Neglected Key to Mitigating Diversity Collapse in Reinforcement Learning with Verifiable Reward' 논문에 대한 자세한 리뷰입니다.#Review#Reinforcement Learning#Large Language Models (LLMs)#Diversity Collapse#f-divergence#Forward-KL#JS-divergence#Pass@k#Catastrophic Forgetting2025년 9월 12일댓글 수 로딩 중