mlfoundations-dev/filter_math_neg_majority_math

该数据集包含了来源信息、标识符、对话内容和问题答案字符串等字段。它被分为训练集，共有37499个示例，数据集大小为1,114,252,832字节。数据集的下载大小为475,565,495字节，并提供了默认配置文件来指定训练集的数据文件。

The dataset includes fields such as source information, identifier, conversation content, and question-answer strings. It is split into a training set with a total of 37,499 examples and a dataset size of 1,114,252,832 bytes. The download size of the dataset is 475,565,495 bytes, and a default configuration file is provided to specify the data files for the training set.