mlsanityData to 14 Oct 2026

Neural Message-Passing on Attention Graphs for Hallucination Detection

arXiv 2509.24770 ↗First posted 29 Sep 2025cs.LGcited by 6

Not on the map. Why?

Abstract (arXiv)

Large Language Models (LLMs) often generate incorrect or unsupported content, known as hallucinations. Existing detection methods rely on heuristics or simple models over isolated computational traces such as activations, or attention maps. We unify these signals by representing them as attributed graphs, where tokens are nodes, edges follow attentional flows, and both carry features from attention scores and activations. Our approach, CHARM, casts hallucination detection as a graph learning task and tackles it by applying GNNs over the above attributed graphs. We show that CHARM provably subsumes prior attention-based heuristics and, experimentally, it consistently outperforms other leading approaches across diverse benchmarks. Our results shed light on the relevant role played by the graph structure and on the benefits of combining computational traces, whilst showing CHARM exhibits promising zero-shot performance on cross-dataset transfer.

Key facts. Cited by 6 papers in the map (data through 14 Oct 2026). Builds on 48 of its 59 references that are in the map. First posted on arXiv 29 Sep 2025.

Builds on

48 of 59 references.

References · 48

  1. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  3. 2016Semi-Supervised Classification with Graph Convolutional Networkscited by 11,361
  4. 2017Neural Message Passing for Quantum Chemistrycited by 3,487
  5. 2019Fast Graph Representation Learning with PyTorch Geometriccited by 2,434
  6. 2022Survey of Hallucination in Natural Language Generationcited by 2,075
  7. 2022Language Models (Mostly) Know What They Knowcited by 1,856
  8. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,601
  9. 2018Relational inductive biases, deep learning, and graph networkscited by 1,360
  10. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  11. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,302
  12. 2017Get To The Point: Summarization with Pointer-Generator Networkscited by 1,292
36 more
  1. 2021Probing Classifiers: Promises, Shortcomings, and Advancescited by 876
  2. 2023Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generationcited by 819
  3. 2023SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Modelscited by 810
  4. 2022Discovering Latent Knowledge in Language Models Without Supervisioncited by 767
  5. 2023The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasetscited by 712
  6. 2018Gender Bias in Coreference Resolution: Evaluation and Debiasing Methodscited by 659
  7. 2023The Internal State of an LLM Knows When It's Lyingcited by 659
  8. 2020On the Bottleneck of Graph Neural Networks and its Practical Implicationscited by 631
  9. 2021Understanding over-squashing and bottlenecks on graphs via curvaturecited by 478
  10. 2024INSIDE: LLMs' Internal States Retain the Power of Hallucination Detectioncited by 323
  11. 2021Understanding Factuality in Abstractive Summarization with FRANK: A Benchmark for Factuality Metricscited by 287
  12. 2024LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinationscited by 244
  13. 2023A Stitch in Time Saves Nine: Detecting and Mitigating Hallucinations of LLMs by Validating Low-Confidence Generationcited by 206
  14. 2024Lookback Lens: Detecting and Mitigating Contextual Hallucinations in Large Language Models Using Only Attention Mapscited by 142
  15. 2021A Token-level Reference-free Hallucination Detection Benchmark for Free-form Text Generationcited by 138
  16. 2023The Troubling Emergence of Hallucination in Large Language Models -- An Extensive Definition, Quantification, and Prescriptive Remediationscited by 131
  17. 2021Hallucinated but Factual! Inspecting the Factuality of Hallucinations in Abstractive Summarizationcited by 129
  18. 2019Fake News Detection on Social Media using Geometric Deep Learningcited by 118
  19. 2023Enhancing Uncertainty-Based Hallucination Detection with Stronger Focuscited by 99
  20. 2024Transformers need glasses! Information over-squashing in language taskscited by 91
  21. 2023Attention Satisfies: A Constraint-Satisfaction Lens on Factual Errors of Language Modelscited by 67
  22. 2022Looking for a Needle in a Haystack: A Comprehensive Study of Hallucinations in Neural Machine Translationcited by 61
  23. 2023On Early Detection of Hallucinations in Factual Question Answeringcited by 56
  24. 2023Detecting and Mitigating Hallucinations in Multilingual Summarisationcited by 48
  25. 2024Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimensioncited by 45
  26. 2023Understanding and Detecting Hallucinations in Neural Machine Translation via Model Introspectioncited by 44
  27. 2019Learning representations of irregular particle-detector geometry with distance-weighted graph networkscited by 43
  28. 2024Benchmarking Hallucination in Large Language Models based on Unanswerable Math Word Problemcited by 43
  29. 2025Hallucination Detection in LLMs Using Spectral Features of Attention Mapscited by 43
  30. 2023Look Before You Leap: An Exploratory Study of Uncertainty Measurement for Large Language Modelscited by 41
  31. 2023Weakly Supervised Detection of Hallucinations in LLM Activationscited by 20
  32. 2024Constructing Benchmarks and Interventions for Combating Hallucinations in LLMscited by 19
  33. 2025Hallucination Detection in LLMs with Topological Divergence on Attention Graphscited by 19
  34. 2025Towards Mechanistic Interpretability of Graph Transformers via Attention Graphscited by 13
  35. 2025Beyond Next Token Probabilities: Learnable, Fast Detection of Hallucinations and Data Contamination on LLM Output Distributionscited by 12
  36. 2025What makes a good feedforward computational graph?cited by 5

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026HALT: Hallucination Assessment via Log-probs as Time seriescited by 8
  2. 2026TDGNet: Hallucination Detection in Diffusion Language Models via Temporal Dynamic Graphscited by 8
  3. 2026Attention Sinks as Internal Signals for Hallucination Detection in Large Language Modelscited by 0
  4. 2026A Graph Signal Processing Perspective on Numerical Sequence Representations in LLM In-Context Learningcited by 0
  5. 2026UniProbe: A Learnable Token-Level Hallucination Detector for Large VLMs using Multi-Structural Internal Representationscited by 0
  6. 2026Gray-Box Hallucination Detection in LLMs with Token Identitiescited by 0