Rubric-based reinforcement learning grades a response against many written criteria but collapses them into one reward, which discards the detail the rubric …
机构:阿里
来源:arXiv 2610.17623 | AI4Papers 论文推荐平台