الملخص
In this work, we examine the topology of information flow patterns within attention graphs to effectively distinguish hallucinated from non-hallucinated responses. We analyze the Forman-Ricci curvature to identify structural patterns indicating information bottlenecks in attention graphs. We then introduce a method that captures both semi-local and global information-flow characteristics of attention heads associated with hallucinated responses. We evaluate our approach extensively across several LLMs and established benchmarks. Empirical results demonstrate that our proposed single-pass approach provides consistent improvements over existing attention-based and multi-response baselines across two hallucination-detection benchmarks, while achieving competitive performance across diverse LLM architectures. Further analysis reveals that impaired context sharing among tokens during causal generation is strongly associated with hallucination occurrences in LLMs. In particular, hallucinated responses are consistently characterized by an over-reliance on self-attention, diffused context retrieval from earlier tokens, or information over-squashing, especially in the final transformer layer.
الكلمات المفتاحية
الموضوع
بيانات النشر
- المجلة
- غير متاح
- وصول مفتوح
- وصول مفتوح أخضر
اقتبس هذه المقالة
APA 7
Jalilifard, A., Rocha, A., Wong, E., & Raimundo, M. M. (2026). Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing. https://omanscience.com/ar/articles/detecting-hallucination-in-llms-tracing-the-topological-signatures-of-impaired-context-sharing
MLA 9
Jalilifard, Amir, et al. "Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing." https://omanscience.com/ar/articles/detecting-hallucination-in-llms-tracing-the-topological-signatures-of-impaired-context-sharing.
شيكاغو (المؤلف–التاريخ)
Jalilifard, Amir, Anderson Rocha, Eric Wong, and Marcos Medeiros Raimundo. 2026. "Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing." https://omanscience.com/ar/articles/detecting-hallucination-in-llms-tracing-the-topological-signatures-of-impaired-context-sharing.
هارفارد
Jalilifard, A., Rocha, A., Wong, E. and Raimundo, M. M. (2026) 'Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing', Available at: https://omanscience.com/ar/articles/detecting-hallucination-in-llms-tracing-the-topological-signatures-of-impaired-context-sharing.
فانكوفر
Jalilifard A, Rocha A, Wong E, Raimundo MM. Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing. https://omanscience.com/ar/articles/detecting-hallucination-in-llms-tracing-the-topological-signatures-of-impaired-context-sharing
IEEE
A. Jalilifard, A. Rocha, E. Wong, and M. M. Raimundo, "Detecting Hallucination in LLMs: Tracing the Topological Signatures of Impaired Context Sharing," https://omanscience.com/ar/articles/detecting-hallucination-in-llms-tracing-the-topological-signatures-of-impaired-context-sharing.