Test-time reinforcement learning adapts a model on its own unlabeled test set using majority-vote pseudo-labels and has shown strong results in mathematics. …
机构:Harvard Medical School
来源:arXiv 2609.16660 | AI4Papers 论文推荐平台