PoLLMgraph: Unraveling Hallucinations in Large Language Models via State Transition Dynamics

Derui Zhu, Dingfan Chen, Qing Li, Zongxiong Chen, Lei Ma, Jens Grossklags, Mario Fritz


Abstract
Despite tremendous advancements in large language models (LLMs) over recent years, a notably urgent challenge for their practical deployment is the phenomenon of "hallucination”, where the model fabricates facts and produces non-factual statements. In response, we propose PoLLMgraph—a Polygraph for LLMs—as an effective model-based white-box detection and forecasting approach. PoLLMgraph distinctly differs from the large body of existing research that concentrates on addressing such challenges through black-box evaluations. In particular, we demonstrate that hallucination can be effectively detected by analyzing the LLM’s internal state transition dynamics during generation via tractable probabilistic models. Experimental results on various open-source LLMs confirm the efficacy of PoLLMgraph, outperforming state-of-the-art methods by a considerable margin, evidenced by over 20% improvement in AUC-ROC on common benchmarking datasets like TruthfulQA. Our work paves a new way for model-based white-box analysis of LLMs, motivating the research community to further explore, understand, and refine the intricate dynamics of LLM behaviors.
Anthology ID:
2024.findings-naacl.294
Volume:
Findings of the Association for Computational Linguistics: NAACL 2024
Month:
June
Year:
2024
Address:
Mexico City, Mexico
Editors:
Kevin Duh, Helena Gomez, Steven Bethard
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
4737–4751
Language:
URL:
https://aclanthology.org/2024.findings-naacl.294
DOI:
Bibkey:
Cite (ACL):
Derui Zhu, Dingfan Chen, Qing Li, Zongxiong Chen, Lei Ma, Jens Grossklags, and Mario Fritz. 2024. PoLLMgraph: Unraveling Hallucinations in Large Language Models via State Transition Dynamics. In Findings of the Association for Computational Linguistics: NAACL 2024, pages 4737–4751, Mexico City, Mexico. Association for Computational Linguistics.
Cite (Informal):
PoLLMgraph: Unraveling Hallucinations in Large Language Models via State Transition Dynamics (Zhu et al., Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-naacl.294.pdf
Copyright:
 2024.findings-naacl.294.copyright.pdf