TAUR-dev/syncotv2_gpt4omini

该数据集包含了问题和解决方案的文本，以及用于评估的多种特征，如cot类型、源类型、元数据、思考轨迹、尝试次数、评分和评分原因等。此外，还包括了输入、输出和总令牌数，成本，索引，模型名称，源数据集，问题列和答案列。数据集分为训练集，其大小为195557字节，包含3个示例。

The dataset contains texts of questions and solutions, along with various features for evaluation such as cot type, source type, metadata, thinking trajectories, attempt counts, grades, and grade reasons. It also includes input, output, and total token counts, cost, index, model name, source dataset, problem column, and answer column. The dataset is split into a training set, which is 195557 bytes in size and contains 3 examples.