Rendra86318/vision-arena-bench-v0.1

VisionArena-Bench是一个包含500个多样化用户提示的自动评估基准，用于通过视觉语言模型（VLM）作为评委来近似Chatbot Arena的模型排名。数据集包含问题ID、主题簇名称、问题提示内容和对应的图像列表。数据集主要用于衡量人类偏好，包含大量STEM相关问题和OCR任务，但较少涉及STEM以外的专业领域问题。数据集的使用需遵守严格的许可协议，包括安全与节制、法律合规等条款。

VisionArena-Bench is an automatic benchmark of 500 diverse user prompts that can be used to cheaply approximate Chatbot Arena model rankings via automatic benchmarking with VLM as a judge. The dataset includes question IDs, cluster names, question prompts, and corresponding image lists. It is primarily designed to measure human preferences and contains a large number of STEM-related questions and OCR tasks, but fewer questions related to specialized domains outside of STEM. The use of the dataset is subject to a strict license agreement, including terms on safety and moderation, legal compliance, etc.