SINdex: Semantic INconsistency Index for Hallucination Detection in LLMs

作     者：Abdaljalil, Samir Kurban, Hasan Sharma, Parichit Serpedin, Erchin Atat, Rachad 

作者机构：Electrical and Computer Engineering Texas A&M University College StationTX United States College of Science and Engineering Hamad Bin Khalifa University Doha Qatar Department of Computer Science Luddy School of Informatics BloomingtonIN United States Department of Computer Science and Mathematics Lebanese American University Beirut Lebanon 

出 版 物：《arXiv》 (arXiv)

年 卷 期：2025年

核心收录：

主　　题：Semantics 

摘      要：Large language models (LLMs) are increasingly deployed across diverse domains, yet they are prone to generating factually incorrect outputs—commonly known as hallucinations. Among existing mitigation strategies, uncertainty-based methods are particularly attractive due to their ease of implementation, independence from external data, and compatibility with standard LLMs. In this work, we introduce a novel and scalable uncertainty-based semantic clustering framework for automated hallucination detection. Our approach leverages sentence embeddings and hierarchical clustering alongside a newly proposed inconsistency measure, SINdex, to yield more homogeneous clusters and more accurate detection of hallucination phenomena across various LLMs. Evaluations on prominent open- and closed-book QA datasets demonstrate that our method achieves AUROC improvements of up to 9.3% over state-of-the-art techniques. Extensive ablation studies further validate the effectiveness of each component in our framework. © 2025, CC BY.