RiverDong/ultrafeedback-personalized

该数据集包含多个配置，每个配置具有相似的特征，包括索引、原始索引、提示、选择、拒绝、数据子集、争议性、反转、满足子集、调查选项、用户类型和属性等。数据集分为训练集和测试集，每个配置的数据集大小和下载大小也有所不同。

The dataset contains multiple configurations, each with similar features including index, original index, prompt, chosen, rejected, data subset, controversial, reversed, satisfied subset, survey options, user type, and attributes. The dataset is divided into training and test sets, with varying dataset sizes and download sizes for each configuration.