WorldVQA：评估多模态大语言模型的原子视觉世界知识
WorldVQA: Measuring Atomic World Knowledge in Multimodal Large Language Models
AI总结：
WorldVQA通过评估多模态大语言模型的原子视觉知识，建立衡量其事实性和百科全书广度的基准测试。
摘要：
我们介绍了WorldVQA，一个旨在评估多模态大语言模型（MLLMs）原子视觉世界知识的基准测试。与当前评估不同，后者往往将视觉知识检索与推理混为一谈，WorldVQA将这些能力分离，严格测量模型‘记忆’的内容。该基准测试评估了在分层分类学中对视觉实体进行定位和命名的原子能力，从常见的头类对象到长尾稀有对象。我们期望WorldVQA能够作为视觉事实性的严格测试，从而建立评估当前及下一代前沿模型百科全书广度和幻觉率的标准。
英文摘要：
We introduce WorldVQA, a benchmark designed to evaluate the atomic visual world knowledge of Multimodal Large Language Models (MLLMs). Unlike current evaluations, which often conflate visual knowledge retrieval with reasoning, WorldVQA decouples these capabilities to strictly measure "what the model memorizes." The benchmark assesses the atomic capability of grounding and naming visual entities across a stratified taxonomy, spanning from common head-class objects to long-tail rarities. We expect WorldVQA to serve as a rigorous test for visual factuality, thereby establishing a standard for assessing the encyclopedic breadth and hallucination rates of current and next-generation frontier models.
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。
