mlfoundations-dev/multiple_samples_none_numina_aime

该数据集包含源信息、问题、解决方案、消息内容及其角色、模型响应、原始行索引以及会话信息。消息和会话字段分别记录了交流的内容和参与者角色。数据集划分为训练集，共有4070个示例，大小为138,763,278字节。

The dataset includes source information, problem, solution, message content and role, model response, original row index, and conversation information. The message and conversation fields record the content of the communication and the role of the participants, respectively. The dataset is divided into a training set with a total of 4070 examples and a size of 138,763,278 bytes.