1
Rescaling Confidence: What Scale Design Reveals About LLM Metacognition
LLM置信度评分尺度（0-100）并非中性设计，新研究揭示其对元认知表现的显著影响。
arXiv:2603.09309v2 Announce Type: replace Abstract: Verbalized confidence, in which LLMs report a numerical certainty score, is widely used to estimat…