视觉语言模型在计数任务中严重偏倚，背景信息干扰判断
Vision Language Models are Biased
- 通过测试7个领域计数任务，发现模型依赖常识而非图像细节
- 平均准确率仅17.05%，移除背景后提升至38.14%
- 揭示过度推理会降低准确性，适合关注模型公平性的研究者
大型语言模型从互联网中记忆了海量知识，这虽有助于下游任务，但也可能导致输出偏向错误或偏见。本文测试了主流视觉语言模型（VLMs）在标准客观视觉任务（如计数与识别）中的表现，发现其在包含动物、标志、国际象棋、棋盘游戏、光学错觉和图案网格等7个不同领域的计数任务中，平均准确率仅为17.05%（例如识别Adidas标志的条纹数量）。去除图像背景后，准确率几乎翻倍（提升21.09个百分点），表明上下文视觉线索触发了模型的偏见响应。进一步分析显示，随着推理令牌增加，计数准确率先上升至约40%，随后下降。本工作揭示了VLMs的一种新失败模式，并提出一种人工监督的自动化框架来检测偏见。代码与数据已公开于vlmsarebiased.github.io。
原文摘要 · Abstract (English)
Large language models (LLMs) memorize a vast amount of prior knowledge from the Internet that helps them on downstream tasks but also may notoriously sway their outputs towards wrong or biased answers. In this work, we test how the knowledge about popular subjects hurt the accuracy of vision language models (VLMs) on standard, objective visual tasks of counting and identification. We find that state-of-the-art VLMs are strongly biased (e.g., unable to recognize the 4th stripe has been added to a 3-stripe Adidas logo) scoring an average of 17.05% accuracy in counting (e.g., counting stripes in an Adidas-like logo) across 7 diverse domains from animals, logos, chess, board games, optical illusions, to patterned grids. Removing image backgrounds nearly doubles accuracy (21.09 percentage points), revealing that contextual visual cues trigger these biased responses. Further analysis of VLMs' reasoning patterns shows that counting accuracy initially rises with thinking tokens, reaching ~40%, before declining with excessive reasoning. Our work presents an interesting failure mode in VLMs and a human-supervised automated framework for testing VLM biases. Code and data are available at: vlmsarebiased.github.io.
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。