skdrx/python-dpo-dataset-formatted

该数据集包含三个主要特征：rejected（拒绝）、chosen（选择）和prompt（提示），每个特征都包含content（内容）和role（角色）两个子特征，数据类型均为字符串。数据集分为一个训练集（train），包含2000个样本，总大小为1910538字节。下载大小为386879字节，数据集总大小为1910538字节。配置文件中指定了默认配置，数据文件路径为data/train-*。

The dataset contains three main features: rejected, chosen, and prompt, each of which includes two sub-features: content and role, with the data type being string. The dataset is divided into one training set (train) containing 2000 samples, with a total size of 1910538 bytes. The download size is 386879 bytes, and the total dataset size is 1910538 bytes. The configuration file specifies the default configuration, with the data file path being data/train-*.