Stable Miscalibration in Large Language Models: A Practical View of High-Confidence Errors

A new arXiv paper (2608.13591v1) finds that high-confidence errors in large language models can be stable under small perturbations, indicating stable miscalibration rather than fragile inference. The study combines an output-level audit score and an internal sensitivity probe, showing that self-critical prompting reduces hidden-state sensitivity across layers in three open-weight models, but audit-defined overconfident errors are not more locally sensitive than correct answers.

arXiv:2608.13591v1 Announce Type: new Abstract: High-confidence errors in large language models are often treated as evidence of fragile internal inference. We study a different possibility: stable miscalibration, where a confident wrong answer remains locally stable under small perturbations. We combine two diagnostics: a label-aware output-level audit score that ranks domains by confidence variation and overconfident mistakes under a forced-answer baseline, and an internal sensitivity probe that measures hidden-state movement. On a multi-domain binary factual audit set, this audit score tracks where abstention-aware self-critique reduces decision loss, although direct labeled baselines rank the same gain more strongly. Internally, self-critical prompting consistently reduces hidden-state sensitivity across layers in three open-weight models. This supports prompt-induced local stabilization rather than a purely output-level abstention pattern, but it does not imply calibration: audit-defined overconfident errors are not clearly more locally sensitive than confidently correct answers, so some high-confidence errors may be stable and miscalibrated rather than simply fragile.