m-a-p/COIG-P-CRM

该数据集包含两个主要部分：被拒绝的内容（rejected）和被选择的内容（chosen）。每个部分都包括内容（content）和角色（role）两个子字段。此外，还包含与被拒绝和被选择内容相关的分数字段，但分数字段的数据类型为null，可能表示这些字段在数据集中不存在或未被填充。数据集分为训练集、测试集和验证集，提供了相应的文件大小和示例数量。数据集的总下载大小和总大小也在文件中给出。

The dataset includes two main sections: rejected content and chosen content. Each section consists of two subfields: content and role. Additionally, there are score fields related to the rejected and chosen content, but the data type of these score fields is null, which may indicate that these fields do not exist or are not populated in the dataset. The dataset is split into training, test, and validation sets, with respective file sizes and example counts provided. The total download size and the overall size of the dataset are also given in the file.