selfcorrexp2/llama3_sft_balanced_gen2_augmath_
收藏Hugging Face2025-01-21 更新2025-04-26 收录
下载链接：
https://hf-mirror.com/datasets/selfcorrexp2/llama3_sft_balanced_gen2_augmath_
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含索引（idx）、提示（prompt）、答案序列（answers）、正确答案（gt）、首次奖励（first_rewards）、预测（prediction）和二次奖励（second_rewards）等字段。数据集被划分为训练集（train），共有27200个示例。由于README中未提供具体描述，无法确定数据集的具体内容和用途。
The dataset includes fields such as index (idx), prompt, answer sequence (answers), correct answer (gt), first reward (first_rewards), prediction (prediction), and second reward (second_rewards). The dataset is split into a training set (train) with a total of 27,200 examples. As the README does not provide a specific description, the specific content and purpose of the dataset cannot be determined.
提供机构：
selfcorrexp2
