SeSE: Black-Box Uncertainty Quantification for Large Language Models Based on Structural Information Theory

arXiv:2511.16275v4 Announce Type: replace-cross Abstract: Reliable uncertainty quantification (UQ) is essential for deploying large language models (LLMs) in safety-critical scenarios, as it enables them to abstain from responding when uncertain, thereby avoiding hallucinations, i.e., plausible yet factually incorrect responses. However, while...

Læs hele artiklen hos kilden.