Nautil：LLM调查员结案决策研究

1. 12:00
   arXiv · Computation and Language

   LLM 调查员何时结案：基于证据的关闭决策研究

   论文研究 LLM 调查员在事故、缺陷和宕机调查中何时应结案，提出结案准确性、证据依赖性和结论与缺口质量三项评测，并构建含 731 份审计案例的 Nautil 数据集。微调 9B 模型后，过度断言从 97% 降至 35%，正确且不夸大的结论从 3% 升至 43%；仅奖励结案决策的强化学习将平衡准确率从 69.2 提升到 83.3。

Current heat 6·Comparable peak 10（Oct 5）·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.