Rescaling Confidence: What Scale Design Reveals About LLM Metacognition

ArXi:2603.09309v1 Announce Type: new Verbalized confidence, in which LLMs report a numerical certainty score, is widely used to estimate uncertainty in black-box settings, yet the confidence scale itself (typically 0--100) is rarely examined. We show that this design choice is not neutral. Across six LLMs and three datasets, verbalized confidence is heavily discretized, with than 78% of responses concentrating on just three round-number values.