Ayush-Singh/skywork-clusters

该数据集包含多个特征，主要用于比较和评估不同选项的优劣。每个数据点包含chosen和rejected两个主要部分，分别代表被选中和被拒绝的选项，每个选项包含内容和角色信息。此外，数据集还提供了来源、选择奖励、拒绝奖励、置信度、簇编号和索引等级等信息。数据集被分为五个簇，每个簇的大小和示例数量不同。

This dataset contains multiple features, primarily used for comparing and evaluating the merits of different options. Each data point includes two main parts: chosen and rejected, representing the selected and rejected options respectively, each containing content and role information. Additionally, the dataset provides source, reward for chosen, reward for rejected, confidence, cluster number, and index level information. The dataset is divided into five clusters, with varying sizes and numbers of examples in each cluster.