因果证据：模态冲突幻觉中的注意力头不平衡
Causal Evidence for Attention Head Imbalance in Modality Conflict Hallucination
AI总结：
本文研究了多模态大语言模型在模态冲突中产生幻觉的原因，通过分析注意力头的因果作用，发现驱动幻觉的头部分布更广且权重更大，而抑制幻觉的头部集中在少量重要头部，提出MACI方法在减少幻觉的同时保持准确性。
摘要：
当多模态大语言模型（MLLMs）优先考虑错误的文本前提而非矛盾的视觉证据时，就会出现模态冲突幻觉。为了理解为什么视觉证据在生成过程中无法占据优势，我们从机制角度出发，考察哪些内部组件驱动或阻碍这一失败。我们通过在五个开源MLLMs上进行头部层面的因果分析，识别出两组具有相反因果作用的注意力头：驱动幻觉的头部和抑制幻觉的头部。我们发现一种一致的不对称性：驱动效应更广泛分布且具有更大的总权重，而抑制效应集中在少量重要头部。消融实验进一步证实，这些组在生成过程中产生相反效果：分布驱动影响和局部抑制共同形成不平衡的路由结构，使生成偏向于错误前提。受此发现启发，我们提出了MACI（模态冲突感知因果干预），一种条件干预方法，仅在检测到冲突时抑制因果识别出的驱动幻觉头部。在五个MLLMs上，MACI在MMMC基准测试中实现了最大的幻觉减少，同时在幻觉准确性之间取得了有利的权衡，并能够零样本转移到SCI-SemanticConflict测试。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
