gupta-tanish/Ultrafeedback-SimPO-seed-7

该数据集包含了提示文本(prompt)、提示ID(prompt_id)、选中的内容(chosen)、被拒绝的内容(rejected)以及消息(messages)等字段。数据集分为训练集和测试集，其中训练集又分为偏好训练集(train_prefs)、SFT训练集(train_sft)和生成训练集(train_gen)，测试集分为偏好测试集(test_prefs)、SFT测试集(test_sft)和生成测试集(test_gen)。每个部分包含不同数量的示例和字节数。

The dataset includes fields such as prompt, prompt_id, chosen, rejected, and messages. It is divided into training and testing sets, with the training set further divided into preference training set (train_prefs), SFT training set (train_sft), and generation training set (train_gen), while the test set is divided into preference test set (test_prefs), SFT test set (test_sft), and generation test set (test_gen). Each part contains a different number of examples and byte sizes.