selfcorrexp2/balanced_model_as_rm_2prompt
收藏官方服务：
资源简介：
该数据集包含了索引、提示文本、答案序列、真实标签以及两个奖励序列（分别对应第一次和第二次的奖励）。数据集被划分为训练集，共有5000个示例。数据集的总大小为13,354,296字节，下载大小为5,308,217字节。
The dataset includes index, prompt text, answer sequence, ground truth label, and two reward sequences (for the first and second rewards respectively). The dataset is split into a training set with a total of 5,000 examples. The total size of the dataset is 13,354,296 bytes, with a download size of 5,308,217 bytes.
提供机构：
selfcorrexp2
