MTRE：面向视觉语言模型幻觉检测的多令牌可靠性估计
MTRE: Multi-Token Reliability Estimation for Hallucination Detection in VLMs
AI总结：
针对现有VLM幻觉检测器仅关注首个令牌logit、忽略早期序列丰富信号的问题，提出轻量白盒方法MTRE，聚合前十个令牌logit检测幻觉，在多个基准上准确率和AUROC均显著优于标准方法，达开源VLM幻觉检测新SOTA。
摘要：
视觉语言模型（VLM）如今在许多多模态任务上的表现已可与人类媲美，但它们仍会产生对象幻觉或生成不安全文本。现有的幻觉检测器，例如单令牌线性探测（LP）和PTrue，通常仅分析生成的第一个令牌的对数概率（logit）或仅其得分最高的分量，忽略了早期令牌分布中蕴含的更丰富信号。我们证明，分析早期logit的完整序列有望提供显著更多的诊断信息。我们强调，幻觉可能仅在若干个令牌之后才显现，因为细微的不一致性会随时间累积。通过分析对应幻觉令牌与非幻觉令牌的logit之间的库尔贝克-莱布勒（KL）散度，我们凸显了纳入后续令牌logit以更准确捕捉VLM可靠性动态的重要性。对此，我们提出了多令牌可靠性估计（MTRE），这是一种轻量级白盒方法，通过多令牌对数似然比和自注意力聚合前十个令牌的logit。尽管存在词汇量大、logit序列长带来的挑战，MTRE仍保持高效且易处理。在MAD-Bench、MM-SafetyBench、MathVista以及四个组合几何基准测试中，MTRE的准确率较标准检测方法提升9.4%，AUROC提升14.8%，创下了开源VLM幻觉检测的新最优水平。
英文摘要：
Vision-language models (VLMs) now rival human performance on many multimodal tasks, yet they still hallucinate objects or generate unsafe text. Current hallucination detectors, e.g., single-token linear probing (LP) and PTrue, typically analyze only the logit of the first generated token or just its highest-scoring component, overlooking richer signals embedded within earlier token distributions. We demonstrate that analyzing the complete sequence of early logits potentially provides substantially more diagnostic information. We emphasize that hallucinations may only emerge after several tokens, as subtle inconsistencies accumulate over time. By analyzing the Kullback-Leibler (KL) divergence between logits corresponding to hallucinated and non-hallucinated tokens, we underscore the importance of incorporating later-token logits to more accurately capture the reliability dynamics of VLMs. In response, we introduce Multi-Token Reliability Estimation (MTRE), a lightweight, white-box method that aggregates logits from the first ten tokens using multi-token log-likelihood ratios and self-attention. Despite the challenges posed by large vocabulary sizes and long logit sequences, MTRE remains efficient and tractable. Across MAD-Bench, MM-SafetyBench, MathVista, and four compositional-geometry benchmarks, MTRE achieves a 9.4% gain in accuracy and a 14.8% gain in AUROC over standard detection methods, establishing a new state of the art in hallucination detection for open-source VLMs.
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。
