mlfoundations-dev/instruction_filtering_scale_up_math_base_random_filtering

该数据集包含多个字段，如指令种子(instruction_seed)、来源(source)、推理(reasoning)等，并提供训练集(train)的相关信息，如字节数和示例数。数据集还包含了对话信息。具体应用场景和数据集内容未在README中详细描述。

The dataset includes multiple fields such as instruction_seed, source, reasoning, etc., and provides information about the training set (train) such as byte size and number of examples. The dataset also contains conversation information. The specific application scenario and content of the dataset are not described in detail in the README.