1
Focal Reward: Balanced Reinforcement Learning under Rubric-Based Rewards
平衡强化学习新方法,利用Focal Reward聚焦关键学习信号,提升训练效率与效果。
arXiv:2605.26579v1 Announce Type: new Abstract: The open-ended generation in LLMs usually requires multi-dimensional rubrics to adequately assess qual…