RiverDong/ultrafeedback-p

该数据集包含多个配置，每个配置都有相似的特征，包括索引、原始索引、提示、选择项、拒绝项、数据子集、是否具有争议性、是否反转、满足子集、调查选项、用户类型和属性等。数据集主要用于训练和测试模型，涉及文本生成、选择与拒绝等任务。

This 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 primarily used for training and testing models, involving tasks such as text generation, selection, and rejection.