soowei/DPO-Zephyr-7B-dataset
收藏官方服务：
资源简介：
该数据集包含用于训练和评估对话生成模型的数据，特别是基于偏好学习的模型。数据集包括prompt（提示）、prompt_id（提示ID）、messages（消息列表）、reference_response（参考响应）、chosen（选择的响应）和rejected（拒绝的响应）等字段。数据集分为test_prefs_1和train_prefs_1两个分割，分别用于测试和训练。
This dataset contains data for training and evaluating dialogue generation models, particularly those based on preference learning. The dataset includes fields such as prompt, prompt_id, messages, reference_response, chosen, and rejected. It is divided into two splits, test_prefs_1 and train_prefs_1, for testing and training purposes respectively.
提供机构：
soowei
