RL-ARC：推理引导的不确定性校准框架

1. 12:00
   arXiv · Artificial Intelligence

   RL-ARC：通过推理引导的不确定性校准大型推理模型

   研究者提出 RL-ARC，一种校准感知训练框架，联合利用推理置信度与答案置信度，对正确样本施加推理引导正则化、对错误样本施加过自信惩罚。在 ID 与 OOD 设置下，RL-ARC 改善校准并让推理模型按问题自适应估计置信度，且不明显牺牲推理性能。论文已被 AACL-IJCNLP 2026 接收，arXiv 编号 2610.11352。

Current heat 9·Comparable peak 10（Oct 9）·Comparable change over 24 hours –

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.