selfcorrexp2/numia_augmath_math_type12_merged
收藏Hugging Face2025-01-18 更新2025-04-26 收录
下载链接:
https://hf-mirror.com/datasets/selfcorrexp2/numia_augmath_math_type12_merged
下载链接
链接失效反馈官方服务:
资源简介:
这是一个文本选择数据集,包含选中的文本(chosen_txt)、被拒绝的文本(rejected_txt)、真实标签(gt)、提示文本(prompt)、选择的文本标识(chosen)和被拒绝的文本标识(rejected),以及一个表示选择差异的浮点数(margin)。数据集有训练集(train)共39407个样本,总大小约为443MB。
This is a text selection dataset containing the chosen text (chosen_txt), rejected text (rejected_txt), ground truth label (gt), prompt text (prompt), chosen text indicator (chosen), rejected text indicator (rejected), and a floating-point number representing the selection margin (margin). The dataset has a training set (train) with 39,407 samples, totaling approximately 443MB in size.
提供机构:
selfcorrexp2


