mlfoundations-dev/instruction_filtering_scale_up_math_base_askllm_16K

这是一个包含对话和推理过程的数据集，适用于自然语言处理任务。数据集特征包括指令种子、来源、是否被使用、分类器推理、原始行索引、推理过程、DeepSeek解决方案、最终推理跟踪以及对话内容。数据集分为训练集，共有16000个示例，总大小为836,868,838字节。

This is a dataset containing conversations and reasoning processes, suitable for natural language processing tasks. The dataset features include instruction seed, source, to be used, classifier reasoning, original row index, reasoning process, DeepSeek solution, final reasoning trace, and conversation content. The dataset is split into a training set with a total of 16,000 examples and a size of 836,868,838 bytes.