xDAN2099/xDAN-Self-Alignment-Reward_datasets-1200k

该数据集包含用于训练的数据，主要特征包括被拒绝的对话内容（rejected）和选择的对话内容（chosen），两者都包含内容（content）和角色（role）两个字段。此外，还有被拒绝内容的评分（rejected_score）和选择内容的评分（chosen_score），以及数据来源（source）和一个索引字段（__index_level_0__）。数据集仅包含一个训练集，共有1,254,903个样本，总大小为4,680,935,459字节。

This dataset contains data for training, with main features including rejected dialogue content (rejected) and chosen dialogue content (chosen), both containing fields for content and role. Additionally, there are scores for rejected content (rejected_score) and chosen content (chosen_score), a source field (source), and an index field (__index_level_0__). The dataset includes only a training set with 1,254,903 samples, totaling 4,680,935,459 bytes.