TAUR-dev/linearized-solutions-rule-based

该数据集包含了多个字段，如解决方案(solution)、问题(question)、思考轨迹(gemini_thinking_trajectory和deepseek_thinking_trajectory)、评分(gemini_grade和deepseek_grade)等。数据集被划分为训练集(train)，大小为776,788,360字节，共有1000个示例。数据集的总大小为776,788,360字节，下载大小为159,636,958字节。

The dataset includes multiple fields such as solution, question, thinking trajectory (gemini_thinking_trajectory and deepseek_thinking_trajectory), grades (gemini_grade and deepseek_grade), etc. The dataset is split into a training set (train) with a size of 776,788,360 bytes and a total of 1000 examples. The total size of the dataset is 776,788,360 bytes, and the download size is 159,636,958 bytes.