mlfoundations-dev/mix_base_code_math_t_0
收藏官方服务：
资源简介：
该数据集包含了指令种子(instruction_seed)、来源(source)、推理过程(reasoning)等字段的信息。数据集分为训练集(train)，共有10000个样本。数据集通过对话的形式组织，每个样本可能包含多个对话回合，其中每个回合包含发送者(from)和对话内容(value)。
The dataset includes information fields such as instruction_seed, source, reasoning, etc. The dataset is split into a training set (train) with a total of 10,000 samples. The dataset is organized in the form of conversations, with each sample potentially containing multiple conversation turns, each with a sender (from) and conversation content (value).
提供机构：
mlfoundations-dev
