Causal Evidence for Attention Head Imbalance in Modality Conflict Hallucination

• URL: http://arxiv.org/abs/2605.19250v1
• Date: Tue, 19 May 2026 01:47:53 GMT
• Title: Causal Evidence for Attention Head Imbalance in Modality Conflict Hallucination
• Abstract summary: Modality-conflict hallucination occurs when multimodal large language models prioritize erroneous textual premises over contradictory visual evidence.<n>We identify two groups of attention heads with opposing causal roles: hallucination-driving heads and hallucination-resisting heads.<n>Motivated by this finding, we propose MACI, a conditional intervention that suppresses causally identified hallucination-driving heads only when conflict is detected.
• Score: 19.688878240689423
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Modality-conflict hallucination occurs when multimodal large language models (MLLMs) prioritize erroneous textual premises over contradictory visual evidence. To understand why visual evidence fails to prevail during generation, we take a mechanistic perspective and examine which internal components drive or resist this failure. We perform head-level causal analysis using path patching across five open-source MLLMs and identify two groups of attention heads with opposing causal roles: hallucination-driving heads and hallucination-resisting heads. We find a consistent asymmetry: driving effects are more broadly distributed and carry greater aggregate weight, whereas resisting effects concentrate in a small number of high-importance heads. Ablation experiments further confirm that these groups exert opposing effects during generation: distributed driving influence and localized resistance together form an imbalanced routing structure that biases generation toward the erroneous premise. Motivated by this finding, we propose MACI (Modality-conflict-Aware Causal Intervention), a conditional intervention that suppresses causally identified hallucination-driving heads only when conflict is detected. Across five MLLMs, MACI achieves the largest hallucination reduction among compared inference-time baselines on the MMMC benchmark with a favorable hallucination-accuracy trade-off, and transfers zero-shot to the SCI-SemanticConflict test.

Related papers

• MLLMs Hallucinate when Information Distribution Drifts in Synergy Heads [16.959361439314616]
  Multimodal Large Language Models (MLLMs) often struggle with hallucinations, thus hindering their reliable practical applications.<n>We propose HEAL, Head-lEvel information disentment and caLibration for identifying and mitigating hallucinations.
  arXiv  Detail & Related papers  (2026-09-05T18:01:22Z)
• Role-Break in Attention Heads: Understanding and Detecting Hallucinations in VLMs [49.743008490747094]
  Vision-Language Models (VLMs) remain prone to hallucinations, producing content that is inconsistent with or unsupported by the input image.<n>Under a unified head-level view, we find that hallucination-induced changes manifest as localized deviations from each head's faithful contextual behavior.<n> Detailed analysis reveals that these deviations are systematically organized across attention heads, contextual sources, and deviation directions, and that the resulting signal is linearly readable once head identity is preserved.
  arXiv  Detail & Related papers  (2026-07-31T13:33:51Z)
• HIVE: Understanding Post-Hallucination Reasoning in Vision Language Models [70.67112531913669]
  Hallucinations in vision language models are commonly treated as semantic errors.<n>Prior work mainly focuses on detecting or suppressing hallucinations at generation time.<n>We study Post Hallucination Reasoning (PHR), the stage in which hallucinated semantics enter the model's inference context.
  arXiv  Detail & Related papers  (2026-07-08T15:02:11Z)
• Local Intrinsic Dimension Unveils Hallucinations in Diffusion Models [18.55564482332065]
  Diffusion models are prone to generating structural hallucinations.<n>We propose a complementary perspective that treats hallucinations as instabilities on the model-induced manifold.<n>We show that a hallucination filter based on such instabilities matches or exceeds the performance of the recently proposed temporal one.
  arXiv  Detail & Related papers  (2026-05-06T15:22:52Z)
• Mitigating Entangled Steering in Large Vision-Language Models for Hallucination Reduction [49.96701537295129]
  Large Vision-Language Models (LVLMs) have achieved remarkable success across cross-modal tasks but remain hindered by hallucinations.<n>Existing methods mitigate hallucinations but often alter generation behavior, resulting in shorter outputs and shifted token distributions.<n>We propose MESA, an effective plug-and-play framework that performs controlled and selective latent intervention for hallucination mitigation.
  arXiv  Detail & Related papers  (2026-04-09T07:31:27Z)
• Exploring Causal Effect of Social Bias on Faithfulness Hallucinations in Large Language Models [50.18087419133284]
  Large language models (LLMs) have achieved remarkable success in various tasks, yet they remain vulnerable to faithfulness hallucinations.<n>We investigate whether social bias contributes to these hallucinations, a causal relationship that has not been explored.<n>A key challenge is controlling confounders within the context, which complicates the isolation of causality between bias states and hallucinations.
  arXiv  Detail & Related papers  (2025-08-11T08:34:28Z)
• Robust Multimodal Large Language Models Against Modality Conflict [94.12341487880465]
  multimodal large language models (MLLMs) are prone to hallucinations in real-world scenarios.<n>We study the inherent conflicts in inputs from different modalities that place MLLMs in a dilemma and directly lead to hallucinations.<n>Three methods are proposed to alleviate the hallucination caused by modality conflict.
  arXiv  Detail & Related papers  (2025-07-09T11:18:38Z)
• MIRAGE: Assessing Hallucination in Multimodal Reasoning Chains of MLLM [58.2298313720146]
  Multimodal hallucinations are multi-sourced and arise from diverse causes.<n>Existing benchmarks fail to adequately distinguish between perception-induced hallucinations and reasoning-induced hallucinations.
  arXiv  Detail & Related papers  (2025-05-30T05:54:36Z)
• Shadows in the Attention: Contextual Perturbation and Representation Drift in the Dynamics of Hallucination in LLMs [6.190663515080656]
  We present the first systematic study linking hallucination incidence to internal-state drift induced by context injection.<n>Using TruthfulQA, we construct two 16-round "titration" tracks per question.<n>We track overt hallucination rates with a tri-perspective detector and covert dynamics via cosine, entropy, JS and Spearman drifts of hidden states and attention maps.
  arXiv  Detail & Related papers  (2025-05-22T16:50:58Z)
• 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)

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.