New methods advance reinforcement learning and safety alignment in large language models

Recent research introduces innovative techniques to improve reinforcement learning with verifiable rewards (RLVR) and preserve safety alignment in large language models (LLMs).