cfpark00/math_linearized_backtracking

该数据集是一个包含问题、答案和相关属性（如难度级别、问题类型等）的数据集，适用于训练和评估模型。数据集分为多个训练集，每个训练集根据问题的随机性或最近问题的正确性进行了不同的划分，并且每个训练集都有小规模和大规模版本。

This dataset is a collection of problems, answers, and associated attributes (such as difficulty level, problem type, etc.) suitable for training and evaluating models. The dataset is divided into multiple training sets, each split according to the randomness of the problems or the correctness of the most recent problems, and each training set comes in both a small-scale and a large-scale version.