RyanYr/DeepSeek-R1-Distill-Qwen-7B_matheval_tokenized

该数据集包含了一系列字段，用于描述数据源、问题、解决方案、答案等信息。它还包括提示内容的详细描述，如内容和角色，以及奖励模型的地面真实和风格。此外，数据集还包含了响应、正确答案、提取的解决方案、奖励模型分数和多种准确度指标。数据集分为训练集，并提供了相应的配置文件。

The dataset comprises multiple fields that describe the data source, problem, solution, answer, detailed prompt content such as content and role, as well as the ground truth and style of the reward model. Additionally, it includes responses, correct answers, extracted solutions, reward model scores, and various accuracy metrics. The dataset is split into a training set and comes with a configuration file specifying the data file paths and split method.