Dataset MCQ LLM 09 25

This dataset contains the complete set of 480 validated MCQs used in the study, the official faculty answers, raw responses from GPT-5 Pro, Gemini 2.5 Pro, and DeepSeek V3, as well as the temporal and contextual variables used in all analyses. Each row corresponds to one MCQ administered to all three models.本数据集包含本研究中使用的全部480道经过验证的多项选择题（Multiple Choice Questions，MCQs）、官方教职人员提供的标准答案，以及GPT-5 Pro、Gemini 2.5 Pro与DeepSeek V3三款模型的原始作答结果，同时涵盖了所有分析所用的时序变量与上下文变量。每一行对应一道向全部三款模型施测的多项选择题。