SiweiWu/mixed_data

这是一个包含提示信息和相关数据的训练数据集，用于训练模型理解不同角色和场景下的文本内容，并作出相应的反应。数据集包含测试集和训练集，每个样本都有多个特征，如数据源、角色、能力、奖励模型的地面真实和风格，以及额外信息如答案、索引、等级、问题和数据集分割信息。

This is a training dataset containing prompt information and related data, designed to train models to understand text content in different roles and scenarios and respond accordingly. The dataset includes test and train sets, with each sample having multiple features such as data source, role, ability, ground truth and style of the reward model, as well as extra information like answer, index, level, question, and dataset split information.