selfcorrexp2/llama3_sft_less_corr_rr60k

该数据集包含多个特征字段，如索引、提示信息、是否为第一轮、正确答案、奖励信息、个人解决方案、标志位、对话轮次以及对话内容。对话内容包括对话文本和角色信息。数据集被划分为训练集，共有241598个示例。数据集的下载大小为594253010字节，实际大小为1660493744字节。

The dataset includes multiple feature fields such as index, prompt, whether its the first round, ground truth, rewards, personal solution, flag, turn, and conversation content. The conversation content includes the text of the conversation and the role information. The dataset is split into a training set with a total of 241598 examples. The download size of the dataset is 594253010 bytes, and the actual size is 1660493744 bytes.