billxbf/ultrafeedback-dpo-iter1

这个数据集包含了对话相关的信息，其中包括提示文本(prompt)、提示ID(prompt_id)、选中的回复(chosen)、被拒绝的回复(rejected)以及对话消息(messages)。每个回复都包括内容和角色信息。此外，数据集中还有一个表示是否交换偏好的布尔字段(swap_preferences)。数据集分为训练集，共有20378个样本，大小为153048459字节。

This dataset includes conversation-related information, such as prompt text (prompt), prompt ID (prompt_id), selected responses (chosen), rejected responses (rejected), and conversation messages (messages). Each response includes content and role information. Additionally, there is a boolean field (swap_preferences) indicating whether preferences are swapped. The dataset is split into a training set with a total of 20378 samples, totaling 153048459 bytes in size.