qfq/cotmath_metamath

该数据集包含1,549,415个训练样本，每个样本包含多个特征字段，如id、sample_id、question_statement、thinking_trajectory、golden_answer、answer、problem、orig_type、orig_problem、orig_original_question、orig_solution和orig_answer等。这些字段描述了样本的详细信息，包括问题陈述、思维轨迹、正确答案、回答、问题类型、原始问题、原始解决方案和原始答案等。数据集的总大小为3,686,327,464字节，下载大小为892,725,665字节。

The dataset contains 1,549,415 training samples, each with multiple feature fields such as id, sample_id, question_statement, thinking_trajectory, golden_answer, answer, problem, orig_type, orig_problem, orig_original_question, orig_solution, and orig_answer. These fields describe detailed information about the samples, including question statements, thinking trajectories, correct answers, responses, problem types, original problems, original solutions, and original answers. The total size of the dataset is 3,686,327,464 bytes, with a download size of 892,725,665 bytes.