[논문리뷰] MSign: An Optimizer Preventing Training Instability in Large Language Models via Stable Rank RestorationarXiv에 게시된 'MSign: An Optimizer Preventing Training Instability in Large Language Models via Stable Rank Restoration' 논문에 대한 자세한 리뷰입니다.#Review#LLM Training Stability#Gradient Explosion#Stable Rank#Jacobian Alignment#Matrix Sign Operation#Optimizer#Transformer2026년 2월 8일댓글 수 로딩 중
[논문리뷰] SR-GRPO: Stable Rank as an Intrinsic Geometric Reward for Large Language Model AlignmentYi Yang이 arXiv에 게시한 'SR-GRPO: Stable Rank as an Intrinsic Geometric Reward for Large Language Model Alignment' 논문에 대한 자세한 리뷰입니다.#Review#LLM Alignment#Stable Rank#Intrinsic Reward#Reinforcement Learning#Geometric Properties#Group Relative Policy Optimization#Annotation-Free Alignment2025년 12월 3일댓글 수 로딩 중