AIPlans/ultrafeedback_binarized

该数据集包含了用于某种任务的用户偏好的对话记录。每个记录包括一个或多个对话消息，每个消息都有内容(content)和角色(role)信息。此外，还包含了提示(prompt)信息、提示ID(prompt_id)、选择的对话(chosen)、被拒绝的对话(rejected)以及对应的选择分数(score_chosen)和拒绝分数(score_rejected)。数据集分为训练集和测试集两部分。

This dataset contains user preference dialogue records for some task. Each record includes one or more dialogue messages, each with content (content) and role (role) information. In addition, it includes prompt (prompt) information, prompt ID (prompt_id), chosen dialogue (chosen), rejected dialogue (rejected), and corresponding choice scores (score_chosen) and rejection scores (score_rejected). The dataset is divided into two parts: a training set and a test set.