selfcorrexp/llama3_prompt_first_corr_math1_processed

该数据集包含多个特征，如索引（idx）、提示（prompt）、答案（answers）、第一轮（first_round）、真实值（gt）、奖励（rewards）、我的解决方案（my_solu）、标志（flag）、轮次（turn）和对话（conversations）。数据集仅包含一个训练集分割，共有80,498个样本，总大小为603,512,030字节，下载大小为208,881,955字节。

The dataset includes various features such as integer-type index, string-type prompt, answers, boolean-type first round flag, string-type ground truth, boolean sequence-type rewards, string sequence-type solution, boolean-type flag, integer-type turn, and a list of conversations containing content and role. The dataset is split into a training set with 80498 samples, with a dataset size of 603512030 bytes and a download size of 208881955 bytes.