raftstudy/pretrained_rm_test_feature

这个数据集包含了选中的(chosen)和被拒绝的(rejected)文本内容及其角色(role)，同时为选中和被拒绝的内容分别提供了特征向量(chosen_feature和rejected_feature)。数据集仅包含一个训练集split，共有7688个示例。

The dataset includes selected (chosen) and rejected (rejected) text content along with their roles (role), and provides feature vectors (chosen_feature and rejected_feature) for both selected and rejected content. The dataset contains only one training set split with a total of 7688 examples.