selfcorrexp2/orm-less-corr-label_llama3_sft_tmp10_vllmexp_rewardtmp07
收藏数据链接：
官方服务：
资源简介：
该数据集包含索引、提示信息、奖励标识、答案序列、真实标签、代理标签和二次奖励标识等字段。数据集分为训练集，共有5000个示例，总大小为14775139字节。提供了一个默认配置，指定了训练集的数据文件路径。
The dataset includes fields such as index, prompt, reward indicator, answer sequence, ground truth label, proxy label, and second reward indicator. The dataset is split into a training set with a total of 5,000 examples and a total size of 14,775,139 bytes. A default configuration is provided, specifying the path to the data files for the training set.
提供机构：
selfcorrexp2
