AI的“学霸秘籍”：原来功夫在“节奏”上

[CL] Bridging Offline and Online Reinforcement Learning for LLMs
[FAIR at Meta]
arxiv.org