HalLoc：视觉语言模型幻觉的Token级定位
HalLoc: Token-level Localization of Hallucinations for Vision Language Models
AI总结：
针对现有视觉语言模型幻觉检测方法计算成本高、输出确定性强的问题，研究提出带15万token级标注的HalLoc数据集及低开销基线模型，可实现概率性并行幻觉检测，提升模型可靠性。
摘要：
幻觉问题对大视觉语言模型的可靠性构成重大挑战，因此检测幻觉对于保障关键应用的准确性至关重要。当前的检测方法通常依赖计算密集型模型，导致高延迟和高资源需求。其确定性的输出结果也无法适配现实场景中幻觉信息与真实信息边界模糊的情况。为解决这些问题，我们提出了HalLoc，这是一个专为高效、概率性幻觉检测设计的数据集。它包含15万个带有token级标注的样本，标注内容涵盖幻觉类型，覆盖视觉问答（VQA）、指令遵循和图像描述任务。该数据集可助力开发具备分级置信度的幻觉检测模型，支持更明智的用户交互。此外，我们提出了一个在HalLoc上训练的基线模型，可在生成过程中以低开销并行执行幻觉检测。该模型可无缝集成到现有视觉语言模型（VLM）中，在保持效率的同时提升可靠性。构建鲁棒的即插即用幻觉检测模块的前景，为提升现实应用中视觉语言模型的可信度开辟了新途径。HalLoc数据集和代码已公开发布于：https://github.com/dbsltm/cvpr25_halloc。
英文摘要：
Hallucinations pose a significant challenge to the reliability of large vision-language models, making their detection essential for ensuring accuracy in critical applications. Current detection methods often rely on computationally intensive models, leading to high latency and resource demands. Their definitive outcomes also fail to account for real-world scenarios where the line between hallucinated and truthful information is unclear. To address these issues, we propose HalLoc, a dataset designed for efficient, probabilistic hallucination detection. It features 150K token-level annotated samples, including hallucination types, across Visual Question Answering (VQA), instruction-following, and image captioning tasks. This dataset facilitates the development of models that detect hallucinations with graded confidence, enabling more informed user interactions. Additionally, we introduce a baseline model trained on HalLoc, offering low-overhead, concurrent hallucination detection during generation. The model can be seamlessly integrated into existing VLMs, improving reliability while preserving efficiency. The prospect of a robust plug-and-play hallucination detection module opens new avenues for enhancing the trustworthiness of vision-language models in real-world applications. The HalLoc dataset and code are publicly available at: https://github.com/dbsltm/cvpr25_halloc.
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。
