SelfCheck-Eval: A multi-module framework for zero-resource hallucination detection in large language models

2026-06-05

0

OA

AI

DOI：10.1016/j.patter.2026.101569

查看原文

摘要

En 中文

• SelfCheck-Eval无需访问大语言模型内部即可检测幻觉 • 引入AIME数学基准用于数学幻觉检测 • 所有现有方法均无法可靠地验证正确的数学推理 • 数学中的幻觉检测需要超越统计学习

Keyword:

hallucination detection

large language models

consistency and uncertainty quantification

context-based verification