mytestdpo/llama3_orm_temp10
收藏Hugging Face2025-01-05 更新2025-04-26 收录
下载链接：
https://hf-mirror.com/datasets/mytestdpo/llama3_orm_temp10
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了一个训练集，其中每个样本都包括一个索引、一个提示、一个答案序列、一个真实标签以及三种类型的奖励（普通奖励、代理奖励和次要奖励）。答案序列和奖励都是序列类型，可能表示与序列预测或生成任务相关。数据集总共包含2632个示例，文件大小为4401400字节。
The dataset consists of a training set where each sample includes an index, a prompt, an answer sequence, a ground truth label, and three types of rewards (regular rewards, proxy rewards, and secondary rewards). The answer sequence and rewards are of sequence type, which may indicate a relation to sequence prediction or generation tasks. The dataset contains a total of 2632 examples with a file size of 4401400 bytes.
提供机构：
mytestdpo
