microsoft/ChatBench

ChatBench数据集包含了用户研究数据和自动化AI实验数据，旨在对同一组MMLU基准问题中AI单独回答、用户与AI结合回答以及用户单独回答进行对比。数据集来源于数学、物理和道德情景等多个领域的问题，并经过了筛选。数据集分为三个部分，分别对应完整研究、两次试点研究，每个部分包含了用户的答案和与AI的对话记录。

ChatBench dataset contains user study data and automated AI-alone experiment data, aiming to compare AI-alone, user-AI, and user-alone answers for the same set of MMLU benchmark questions. The dataset is sourced from questions in various fields such as mathematics, physics, and moral scenarios, and has undergone filtering. The dataset is divided into three parts, corresponding to the full study and two pilot studies, each containing user answers and conversations with AI.