BeliefScope：诊断LLM证据驱动与压力诱导的信念修正

1. 12:00
   arXiv · Computation and Language

   BeliefScope：诊断大语言模型的证据驱动修正与压力诱导偏移

   研究提出 BeliefScope，一种受控黑盒框架，用于在固定目标命题下区分证据与方向性用户压力对模型响应变化的影响。框架将 Evidence 与 Pressure 与因子级局部控制交叉，通过概率报告、类别判断和行动推荐测量响应变化，并在受控合成条件下评估可归因性。

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.