mytestdpo/type1_halftype2_llama3it_gsm8k
收藏Hugging Face2025-01-17 更新2025-04-26 收录
下载链接：
https://hf-mirror.com/datasets/mytestdpo/type1_halftype2_llama3it_gsm8k
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含选择的文本和被拒绝的文本，以及与之相关的标签和提示文本。它还包括一个ground truth字段和一个边缘分数字段，用于表示预测的不确定性。数据集分为训练集，并提供了详细的大小和示例数量信息。
The dataset consists of chosen and rejected texts along with their associated labels and prompt text. It also includes a ground truth field and a margin score field to represent the uncertainty of predictions. The dataset is split into a training set and detailed information about the size and number of examples is provided.
提供机构：
mytestdpo
