# Causal Evidence for Attention Head Imbalance in Modality Conflict Hallucination

arXiv: https://arxiv.org/abs/2605.19250 · HTML: https://mlsanity.com/paper/2605.19250/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Hallucination mitigation in LVLMs, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 18 of its 24 references that are in the map.
- First posted on arXiv 19 May 2026.

## Placed in

- Multimodal LLMs / vision-language models / [Hallucination mitigation in LVLMs](https://mlsanity.com/mllm/hallucination-mitigation-in-lvlms/)

## Builds on (most cited first, 18 of 18)

- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models](https://mlsanity.com/paper/2504.10479/) (2025)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models](https://mlsanity.com/paper/2512.02556/) (2025)
- [Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding](https://mlsanity.com/paper/2311.16922/) (2023)
- [HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Models](https://mlsanity.com/paper/2310.14566/) (2023)
- [OPERA: Alleviating Hallucination in Multi-Modal Large Language Models via Over-Trust Penalty and Retrospection-Allocation](https://mlsanity.com/paper/2311.17911/) (2023)
- [Hallucination of Multimodal Large Language Models: A Survey](https://mlsanity.com/paper/2404.18930/) (2024)
- [Mitigating Hallucinations in Large Vision-Language Models with Instruction Contrastive Decoding](https://mlsanity.com/paper/2403.18715/) (2024)
- [Cross-modal Information Flow in Multimodal Large Language Models](https://mlsanity.com/paper/2411.18620/) (2024)
- [Understanding Information Storage and Transfer in Multi-modal Large Language Models](https://mlsanity.com/paper/2406.04236/) (2024)
- [ASCD: Attention-Steerable Contrastive Decoding for Reducing Hallucination in MLLM](https://mlsanity.com/paper/2506.14766/) (2025)
- [Cracking the Code of Hallucination in LVLMs with Vision-aware Head Divergence](https://mlsanity.com/paper/2412.13949/) (2024)
- [Robust Multimodal Large Language Models Against Modality Conflict](https://mlsanity.com/paper/2507.07151/) (2025)
- [Dissecting Dissonance: Benchmarking Large Multimodal Models Against Self-Contradictory Instructions](https://mlsanity.com/paper/2408.01091/) (2024)
- [Intervene-All-Paths: Unified Mitigation of LVLM Hallucinations across Alignment Formats](https://mlsanity.com/paper/2511.17254/) (2025)
- [Challenges in Understanding Modality Conflict in Vision-Language Models](https://mlsanity.com/paper/2509.02805/) (2025)

## Built on by (most cited first, 1 of 1)

- [Inherited Heads: Audio language models track speakers with their text backbone's attention, and an attention-mass ranking retrieves a different set](https://mlsanity.com/paper/2609.14174/) (2026)

## Abstract (arXiv)

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.