[UAI 2026 Oral] SeSE: Black-Box Uncertainty Quantification for Large Language Models Based on Structural Information Theory, which aims to detect hallucinated content in LLM-generated text.
pythonnlpblack-boxinformation-theorypytorchquestion-answeringuncertainty-quantificationtrustworthy-aillmhallucination-detectionllm-hallucinationsemantic-entropyllm-uncertaintystructural-entropyuai-2026
-
Updated
Jul 6, 2026 - Python