Rexhaif/hh-rlhf-chat-template

该数据集包含三个主要特征：chosen、rejected和conversation。chosen和rejected特征都包含content和role两个子特征，类型均为字符串。conversation特征是一个列表，列表中的每个元素包含content和role两个子特征，类型同样为字符串。数据集分为训练集和测试集，训练集包含160800个样本，测试集包含8552个样本。数据集的下载大小为127647932字节，总大小为220953333字节。数据集适用于文本生成任务，语言为英语，标签包括dpo、rl和rlhf，数据集大小在100K到1M之间。

The dataset contains three main features: chosen, rejected, and conversation. Both chosen and rejected features include sub-features content and role, both of which are of string type. The conversation feature is a list, where each element in the list includes sub-features content and role, also of string type. The dataset is divided into a training set and a test set, with 160800 samples in the training set and 8552 samples in the test set. The download size of the dataset is 127647932 bytes, and the total size is 220953333 bytes. The dataset is suitable for text generation tasks, in English, with tags including dpo, rl, and rlhf, and the dataset size is between 100K and 1M.