Jennny/randomized_ultrafeedback_prefs

该数据集包含多个字段，如prompt（提示）、prompt_id（提示ID）、chosen（选择的回答）、rejected（拒绝的回答）、messages（消息）等。每个字段都有其特定的数据类型和结构，例如chosen和rejected字段包含content（内容）和role（角色）两个子字段。数据集分为train_prefs和test_prefs两个部分，分别包含165264和6000个示例，总大小约为1.12GB。

The dataset contains multiple fields such as prompt, prompt_id, chosen, rejected, messages, etc. Each field has its specific data type and structure, for example, the chosen and rejected fields include subfields content and role. The dataset is divided into train_prefs and test_prefs, containing 165264 and 6000 examples respectively, with a total size of approximately 1.12GB.