Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associations

• URL: http://arxiv.org/abs/2504.12691v1
• Date: Thu, 17 Apr 2025 06:34:45 GMT
• Title: Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associations
• Authors: Yiyou Sun, Yu Gai, Lijie Chen, Abhilasha Ravichander, Yejin Choi, Dawn Song,
• Abstract summary: This paper introduces a subsequence association framework to systematically trace and understand hallucinations.<n>Key insight is hallucinations that arise when dominant hallucinatory associations outweigh faithful ones.<n>We propose a tracing algorithm that identifies causal subsequences by analyzing hallucination probabilities across randomized input contexts.
• Score: 82.42811602081692
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Large language models (LLMs) frequently generate hallucinations-content that deviates from factual accuracy or provided context-posing challenges for diagnosis due to the complex interplay of underlying causes. This paper introduces a subsequence association framework to systematically trace and understand hallucinations. Our key insight is that hallucinations arise when dominant hallucinatory associations outweigh faithful ones. Through theoretical and empirical analyses, we demonstrate that decoder-only transformers effectively function as subsequence embedding models, with linear layers encoding input-output associations. We propose a tracing algorithm that identifies causal subsequences by analyzing hallucination probabilities across randomized input contexts. Experiments show our method outperforms standard attribution techniques in identifying hallucination causes and aligns with evidence from the model's training corpus. This work provides a unified perspective on hallucinations and a robust framework for their tracing and analysis.

Related papers

• Towards Long Context Hallucination Detection [49.195854802543714]
  Large Language Models (LLMs) have demonstrated remarkable performance across various tasks. They are prone to contextual hallucination, generating information that is either unsubstantiated or contradictory to the given context. We propose a novel architecture that enables pre-trained encoder models, such as BERT, to process long contexts and effectively detect contextual hallucinations.
  arXiv  Detail & Related papers  (2025-04-28T03:47:05Z)
• HalluLens: LLM Hallucination Benchmark [49.170128733508335]
  Large language models (LLMs) often generate responses that deviate from user input or training data, a phenomenon known as "hallucination" This paper introduces a comprehensive hallucination benchmark, incorporating both new extrinsic and existing intrinsic evaluation tasks.
  arXiv  Detail & Related papers  (2025-04-24T13:40:27Z)
• HalCECE: A Framework for Explainable Hallucination Detection through Conceptual Counterfactuals in Image Captioning [5.130890556960832]
  This work delves into the intricacies of hallucinatory phenomena exhibited by widely used image captioners, unraveling interesting patterns.<n>The deterministic and efficient nature of the employed conceptual counterfactuals backbone is able to suggest semantically minimal edits.<n>Our proposed hallucination detection framework is highly interpretable, by providing semantically meaningful edits apart from standalone numbers.
  arXiv  Detail & Related papers  (2025-03-01T10:28:19Z)
• Who Brings the Frisbee: Probing Hidden Hallucination Factors in Large Vision-Language Model via Causality Analysis [14.033320167387194]
  A major challenge in their real-world application is hallucination, where LVLMs generate non-existent visual elements, eroding user trust. We hypothesize that hidden factors, such as objects, contexts, and semantic foreground-background structures, induce hallucination. By analyzing the causality between images, text prompts, and network saliency, we systematically explore interventions to block these factors.
  arXiv  Detail & Related papers  (2024-12-04T01:23:57Z)
• Detecting and Mitigating Hallucination in Large Vision Language Models via Fine-Grained AI Feedback [40.930238150365795]
  We propose detecting and mitigating hallucinations in Large Vision Language Models (LVLMs) via fine-grained AI feedback.<n>We generate a small-size hallucination annotation dataset by proprietary models.<n>Then, we propose a detect-then-rewrite pipeline to automatically construct preference dataset for training hallucination mitigating model.
  arXiv  Detail & Related papers  (2024-04-22T14:46:10Z)
• Alleviating Hallucinations of Large Language Models through Induced Hallucinations [67.35512483340837]
  Large language models (LLMs) have been observed to generate responses that include inaccurate or fabricated information. We propose a simple textitInduce-then-Contrast Decoding (ICD) strategy to alleviate hallucinations.
  arXiv  Detail & Related papers  (2023-12-25T12:32:49Z)
• HalluciDoctor: Mitigating Hallucinatory Toxicity in Visual Instruction Data [102.56792377624927]
  hallucinations inherent in machine-generated data remain under-explored. We present a novel hallucination detection and elimination framework, HalluciDoctor, based on the cross-checking paradigm. Our method successfully mitigates 44.6% hallucinations relatively and maintains competitive performance compared to LLaVA.
  arXiv  Detail & Related papers  (2023-11-22T04:52:58Z)
• AutoHall: Automated Hallucination Dataset Generation for Large Language Models [56.92068213969036]
  This paper introduces a method for automatically constructing model-specific hallucination datasets based on existing fact-checking datasets called AutoHall. We also propose a zero-resource and black-box hallucination detection method based on self-contradiction.
  arXiv  Detail & Related papers  (2023-09-30T05:20:02Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.