通过动态选择注意力头干预，减少大模型视觉语言幻觉。
Dynamic Multimodal Activation Steering for Hallucination Mitigation in Large Vision-Language Models
- 根据语义匹配动态选择注意力头进行干预
- 在多个数据集上显著降低幻觉率，优于现有方法
- 无需训练，适合部署在各类视觉语言模型中
大型视觉语言模型在视觉语言任务中表现优异，但存在幻觉问题。通过对模型激活模式的深入分析，我们发现：1）真实性与视觉感知能力主要激活模型中不同的注意力头子集；2）真实性引导向量在不同语义上下文中差异显著。基于此，我们提出无训练的动态多模态激活引导方法，构建基于语义的真实感引导向量数据库，并计算视觉感知引导向量，在推理时根据输入语义相似度动态选择最相关引导向量，作用于最具影响力的注意力头。我们在多个模型和数据集上进行实验，结果表明该方法显著提升模型性能，优于现有最先进方法。
原文摘要 · Abstract (English)
Large Vision-Language Models (LVLMs) exhibit outstanding performance on vision-language tasks but struggle with hallucination problems. Through in-depth analysis of LVLM activation patterns, we reveal two key findings: 1) truthfulness and visual perception capabilities predominantly engage different subsets of attention heads within the model architecture; and 2) truthfulness steering vectors vary significantly across different semantic contexts. Based on these observations, we propose Dynamic Multimodal Activation Steering, a training-free approach for hallucination mitigation. Our method constructs a semantic-based truthfulness steering vector database and computes visual perception steering vectors, enabling context-aware interventions during inference by dynamically selecting the most relevant steering vectors based on input semantic similarity and applying them to the most influential attention heads. We conduct comprehensive experiments across multiple models and datasets, demonstrating that our approach significantly enhances model performance, outperforming existing state-of-the-art methods.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。