kaiwenw/dec7_aft_llama8b_1.1

该数据集包含五个主要特征：prompt（提示）、chosen（选择的回答）、chosen_score（选择的回答得分）、rejected（拒绝的回答）和rejected_score（拒绝的回答得分）。数据集分为训练集和验证集两个部分，训练集包含3457个样本，验证集包含186个样本。数据集的总下载大小为19093585字节，总数据集大小为28180417字节。The dataset contains five main features: prompt, chosen, chosen_score, rejected, and rejected_score. It is divided into two parts: a training set with 3457 examples and a validation set with 186 examples. The total download size of the dataset is 19093585 bytes, and the total dataset size is 28180417 bytes.