selfcorrexp2/llama3_sft_more_corr_rr60k_orm_training

该数据集包含多个字段，如索引、提示、是否为第一轮、真实标签、奖励、个人解决方案、标志位、对话轮数、对话内容（包括角色和内容）、难度等级、类型、预测结果以及个人提示对话内容等。数据集分为训练集，共有307127个样本，总文件大小为1.5GB。但数据集具体的应用场景和内容描述未在README中给出。

The dataset includes fields such as index, prompt, whether it is the first round, ground truth, rewards, personal solutions, flags, turn number, conversation content (including role and content), difficulty level, type, prediction results, and personal prompt conversation content. The dataset is split into a training set with a total of 307127 samples and a total file size of 1.5GB. However, the specific application scenario and content description of the dataset are not provided in the README.