qfq/s1K_v2_rep3

该数据集包含了问题的解决方案（solution）、问题本身（question）、协作类型（cot_type）、来源类型（source_type）、元数据（metadata）、协作文本（cot，可为空）、思考轨迹（thinking_trajectories，序列字符串）和尝试（attempt）等字段。数据集被划分为训练集，其中包含1000个示例，总大小约为16716317.738711914字节。同时，提供了默认配置，指定了训练数据文件的路径。

The dataset includes fields for solution to the problem, the question itself, type of collaboration (cot_type), source type (source_type), metadata, collaborative text (cot, can be null), thinking trajectories (sequence of strings), and attempts. The dataset is split into a training set with 1000 examples, totaling approximately 16716317.738711914 bytes in size. Additionally, a default configuration is provided, specifying the path to the training data files.