视觉-语言反思：通过可解释的双因果引导减轻大语言模型中的过度自信幻觉
Vision-Language Introspection: Mitigating Overconfident Hallucinations in MLLMs via Interpretable Bi-Causal Steering
AI总结：
VLI通过可解释的双因果引导方法，有效减少大语言模型中的对象幻觉，提升多模态任务的准确性和可靠性。
摘要：
对象幻觉严重损害了多模态大语言模型的可靠性，通常源于认知反思的根本失败，即模型盲目信任语言先验而非特定视觉证据。现有缓解措施仍显有限：对比解码方法仅在表层操作而未能纠正内部语义错位，而当前潜在引导方法依赖静态向量缺乏实例特定精度。我们引入视觉-语言反思（VLI），一种无需训练的推理框架，模拟元认知自我校正过程。VLI首先通过概率冲突检测进行属性反思，诊断幻觉风险并定位因果视觉锚点。然后运用可解释的双因果引导主动调节推理过程，动态隔离视觉证据与背景噪声，通过自适应校准消除盲目自信。VLI在先进模型上取得最佳性能，减少对象幻觉率12.67%在MMHal-Bench上，提高准确性5.8%在POPE上。
英文摘要：
Object hallucination critically undermines the reliability of Multimodal Large Language Models, often stemming from a fundamental failure in cognitive introspection, where models blindly trust linguistic priors over specific visual evidence. Existing mitigations remain limited: contrastive decoding approaches operate superficially without rectifying internal semantic misalignments, while current latent steering methods rely on static vectors that lack instance-specific precision. We introduce Vision-Language Introspection (VLI), a training-free inference framework that simulates a metacognitive self-correction process. VLI first performs Attributive Introspection to diagnose hallucination risks via probabilistic conflict detection and localize the causal visual anchors. It then employs Interpretable Bi-Causal Steering to actively modulate the inference process, dynamically isolating visual evidence from background noise while neutralizing blind confidence through adaptive calibration. VLI achieves state-of-the-art performance on advanced models, reducing object hallucination rates by 12.67% on MMHal-Bench and improving accuracy by 5.8% on POPE.
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。
