[논문리뷰] SafeGRPO: Self-Rewarded Multimodal Safety Alignment via Rule-Governed Policy OptimizationBo Du이 arXiv에 게시한 'SafeGRPO: Self-Rewarded Multimodal Safety Alignment via Rule-Governed Policy Optimization' 논문에 대한 자세한 리뷰입니다.#Review#Multimodal Safety Alignment#Rule-Governed RL#Self-Rewarded Learning#MLLM Safety#Policy Optimization#Safety Benchmarking#Compositional Robustness2025년 11월 17일댓글 수 로딩 중