CDH-Bench：一种基于常识的幻觉基准，用于评估视觉语言模型中的视觉保真度
CDH-Bench: A Commonsense-Driven Hallucination Benchmark for Evaluating Visual Fidelity in Vision-Language Models
AI总结：
本文提出CDH-Bench基准，用于评估视觉语言模型在视觉证据与常识冲突时的视觉保真度，通过三种异常类型和两种QA任务，揭示模型在常识驱动下的幻觉问题。
摘要：
视觉语言模型（VLMs）在许多基准上表现优异，但一个基本可靠性问题仍未被深入探讨：当视觉证据与常识冲突时，模型是遵循所见还是常识？这种情况下的一种典型失败是模型会覆盖视觉证据并输出常识替代。我们将其称为常识驱动幻觉（CDH）。为评估此现象，我们引入CDH-Bench基准，旨在创建明确的视觉证据-常识冲突。CDH-Bench涵盖三个维度：计数异常、关系异常和属性异常。我们评估了前沿VLMs在二元问答（QA）和多选QA任务下的表现，并报告了包括反事实准确率（CF-Acc）、常识准确率（CS-Acc）、反事实准确率下降（CFAD）、常识崩溃率（CCR）和相对先验依赖（RPD）等指标。结果表明，即使强大的模型在视觉证据与常识冲突下仍易受先验驱动的归一化影响。CDH-Bench为在视觉证据与常识冲突下的视觉保真度提供了受控诊断。
英文摘要：
Vision-language models (VLMs) achieve strong performance on many benchmarks, yet a basic reliability question remains underexplored: when visual evidence conflicts with commonsense, do models follow what is shown or what commonsense suggests? A characteristic failure in this setting is that the model overrides visual evidence and outputs the commonsense alternative. We term this phenomenon \textbf{commonsense-driven hallucination} (CDH). To evaluate it, we introduce \textbf{CDH-Bench}, a benchmark designed to create explicit \textbf{visual evidence--commonsense conflicts}. CDH-Bench covers three dimensions: \textit{counting anomalies}, \textit{relational anomalies}, and \textit{attribute anomalies}. We evaluate frontier VLMs under \textit{binary Question Answering (QA)} and \textit{multiple-choice QA}, and report metrics including \textit{Counterfactual Accuracy} (CF-Acc), \textit{Commonsense Accuracy} (CS-Acc), \textit{Counterfactual Accuracy Drop} (CFAD), \textit{Commonsense Collapse Rate} (CCR), and \textit{Relative Prior Dependency} (RPD). Results show that even strong models remain vulnerable to prior-driven normalization under visual evidence--commonsense conflict. CDH-Bench provides a controlled diagnostic of visual fidelity under visual evidence--commonsense conflict.
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。
