selfcorrexp2/llama3_sft_star_data

这是一个包含对话和回答的数据集，具体包含字段有索引、提示信息、是否为第一轮对话、真实答案、奖励信息、我的解决方案、标志位、对话轮数以及对话内容。对话内容包括对话文本和角色信息。数据集分为训练集，大小为967489469字节，共有124654个示例。数据集下载大小为359668843字节。

This is a dataset containing dialogues and responses, which includes fields such as index, prompt information, whether it is the first round of the dialogue, the ground truth answer, reward information, my solution, flag, dialogue turn, and dialogue content. The dialogue content includes the text of the dialogue and the role information. The dataset is split into a training set, which is 967489469 bytes in size and contains 124654 examples. The download size of the dataset is 359668843 bytes.