arXiv · NLP· Andrew Perrault (Relantic, Inc., Columbus, OH, USA), Anmol Takiar (Relantic, Inc., Columbus, OH, USA), Dominik Winecki (Relantic, Inc., Columbus, OH, USA), Arnab Nandi (Relantic, Inc., Columbus, OH, USA)·· 2026-10-14AI 评分29
RealityBench:评估视觉语言模型记录人类体力工作的模拟环境基准
RealityBench: Simulated Environments to Evaluate a Vision Language Model's Ability to Document Human Physical Work
AI 导读
研究者构建RealityBench基准,利用Blender模拟11种非结构化物理工作环境,评估26个视觉语言模型。模型在文档记录上表现良好,但难以准确判断图像的可接受性(如证明工作的图片质量)。
来源:arXiv · NLP · arxiv.org