FlippyDora/dpo_pair_data_filtered_length15
收藏官方服务：
资源简介：
该数据集包含用于训练和评估模型的多项特征，包括提示（prompt）、选择的回答（chosen）、拒绝的回答（rejected）以及相应的评分（chosen_score和rejected_score）。数据集还记录了训练步骤（step）和总步骤数（num_steps）。数据集被组织为一个训练集，包含41217个样本，文件大小约为47.9MB。
This dataset includes multiple features for training and evaluating models, such as prompts, chosen responses, rejected responses, and corresponding scores (chosen_score and rejected_score). It also records training steps (step) and the total number of steps (num_steps). The dataset is organized into a training set containing 41,217 samples, with a file size of approximately 47.9MB.
提供机构：
FlippyDora
