tttx/3k-forcing-022225-L50-collate

该数据集包含对话信息，每个对话由内容(content)和角色(role)组成。除此之外，每个样本还包含难度(difficulty)级别、问题唯一标识符(problem_uid)和步骤(step)信息。数据集分为训练集和测试集，共有400个训练样本和1个测试样本。

The dataset consists of conversation information, with each conversation comprising content and role. In addition, each sample also includes difficulty level, problem unique identifier, and step information. The dataset is split into a training set and a test set, containing 400 training samples and 1 test sample respectively.