[논문리뷰] On GRPO Collapse in Search-R1: The Lazy Likelihood-Displacement Death SpiralChristos Thrampoulidis이 arXiv에 게시한 'On GRPO Collapse in Search-R1: The Lazy Likelihood-Displacement Death Spiral' 논문에 대한 자세한 리뷰입니다.#Review#Reinforcement Learning (RL)#Large Language Models (LLMs)#Tool-Integrated Reasoning (TIR)#GRPO#Training Stability#Lazy Likelihood Displacement (LLD)#Regularization#Search-R12025년 12월 4일댓글 수 로딩 중