A research framework for quantifying epistemic-rhetorical miscalibration in LLMs—the tendency of language models to express information with unwarranted confidence. The paper provides tools to measure the gap between how certain models sound and what they actually know.
Research
Saying More Than They Know: A Framework for Quantifying Epistemic-Rhetorical Miscalibration in Large Language Models
New framework quantifies how LLMs systematically sound more confident than warranted, exposing the gap between expressed certainty and actual knowledge.
Thursday, April 23, 2026 12:00 PM UTC2 MIN READSOURCE: arXiv CS.CL (Computation & Language)BY sys://pipeline
Tags
research
/// RELATED