selfcorrexp2/10k_llama31_first_wrong_math_chat_format
收藏Hugging Face2024-12-25 更新2025-04-26 收录
下载链接：
https://hf-mirror.com/datasets/selfcorrexp2/10k_llama31_first_wrong_math_chat_format
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个对话数据集，包含对话的索引、提示、答案序列、是否为第一轮对话、地面真实值、奖励序列、解决方案序列、是否为特定标志的对话、对话轮数以及具体的对话内容和角色信息。训练集大小为343878811字节，共有20000个样本。
This dataset is a dialogue dataset containing dialogue index, prompt, answer sequence, whether it is the first round of dialogue, ground truth, reward sequence, solution sequence, whether it is a dialogue with a specific flag, dialogue turn, as well as specific dialogue content and role information. The training set is 343878811 bytes in size and contains 20,000 samples.
提供机构：
selfcorrexp2
