SiweiWu/mixed_data_v2

该数据集包含文本信息，其中包括数据源、提示内容（包含文本内容和角色）、能力描述、奖励模型（包括地面真实和风格）、以及额外信息（包括答案、索引、等级、问题和分割信息）。数据集分为测试集和训练集，测试集包含2616个示例，训练集包含58686个示例。

The dataset includes text information such as data source, prompt content (including text content and role), ability description, reward model (including ground truth and style), and extra information (including answer, index, level, question, and split). The dataset is split into test and train sets, with the test set containing 2616 examples and the train set containing 58686 examples.