japan-ai-official/jmmlu-curated-math

该数据集包含了一系列的问题和多个选项（A、B、C、D），每个问题都标记了正确答案。此外，数据集中还包含了问题的科目、模型的预测结果、问题提示、样本的唯一标识符和问题的难度分数。数据集专为训练机器学习模型设计，提供了训练集 split。数据集大小为193487字节，共有227个训练样本。

The dataset consists of a series of questions along with multiple choices (A, B, C, D) for each question, each with a labeled correct answer. Additionally, the dataset includes the subject of the question, the models prediction, a prompt for the question, a unique identifier for each sample, and a difficulty score for the question. The dataset is designed for training machine learning models and provides a training set split. The dataset size is 193487 bytes, with a total of 227 training samples.