Leveraging Graph Structures to Detect Hallucinations in Large Language Models

Open Access
Authors
Publication date 2024
Host editors
  • D. Ustalov
  • Y. Gao
  • A. Panchenko
  • E. Tutubalina
  • I. Nikishina
  • A. Ramesh
  • A. Sakhovskiy
  • R. Usbeck
  • G. Penn
  • M. Valentino
Book title Proceedings of TextGraphs-17: Graph-based Methods for Natural Language Processing : The 62nd Annual Meeting of the Association of Computational Linguistics
Book subtitle TextGraphs @ ACL 2024 : August 15, 2024
ISBN (electronic)
  • 9798891761452
Event TextGraphs-17: Graph-based Methods for Natural Language Processing
Pages (from-to) 93-104
Publisher Kerrville, TX: Association for Computational Linguistics
Organisations
  • Faculty of Science (FNWI) - Informatics Institute (IVI)
Abstract
Large language models are extensively applied across a wide range of tasks, such as customer support, content creation, educational tutoring, and providing financial guidance. However, a well-known drawback is their predisposition to generate hallucinations. This damages the trustworthiness of the information these models provide, impacting decision-making and user confidence. We propose a method to detect hallucinations by looking at the structure of the latent space and finding associations within hallucinated and non-hallucinated generations. We create a graph structure that connects generations that lie closely in the embedding space. Moreover, we employ a Graph Attention Network which utilizes message passing to aggregate information from neighboring nodes and assigns varying degrees of importance to each neighbor based on their relevance. Our findings show that 1) there exists a structure in the latent space that differentiates between hallucinated and non-hallucinated generations, 2) Graph Attention Networks can learn this structure and generalize it to unseen generations, and 3) the robustness of our method is enhanced when incorporating contrastive learning. When evaluated against evidence-based benchmarks, our model performs similarly without access to search-based methods.
Document type Conference contribution
Language English
Published at https://doi.org/10.18653/v1/2024.textgraphs-1.7
Downloads
2024.textgraphs-1.7 (Final published version)
Permalink to this page
Back