japan-ai-official/jmmlu-curated-humanities

该数据集包含问题及其四个选项（A、B、C、D），正确答案，所属科目，模型的预测结果，问题提示，样本的唯一标识符以及难度评分。这些信息被用于训练和评估可能用于教育或考试评分的模型。数据集分为训练集，包含293个示例，总大小为345,745字节。

The dataset includes questions with four options (A, B, C, D), the correct answer, the subject, the models prediction, a prompt for the question, a unique identifier for the sample, and a difficulty score. This information is used for training and evaluating models that might be used for educational or exam scoring. The dataset is split into a training set with 293 examples, totaling 345,745 bytes in size.