gabrielmbmb/math-500-dota-math

该数据集包含500个示例，主要用于数学问题的变体训练，特别是关于将直角坐标转换为极坐标的问题。数据集通过distilabel工具生成，并包含一个`pipeline.yaml`文件，用于复现生成数据集的流程。数据集的特征包括查询、增强查询、distilabel元数据和模型名称。元数据中包含了输入和输出的token统计信息。

The math-500-dota-math dataset is a synthetic dataset created using the Distilabel tool. It contains a pipeline configuration that can be used to reproduce the dataset generation process. The dataset includes features such as query, augmented_queries, distilabel_metadata, and model_name. The augmented_queries feature contains a list of queries that have been augmented in various ways, such as introducing fractions, combining multiple concepts, and increasing problem complexity. The distilabel_metadata feature includes metadata about the raw input and output, as well as statistics about the tokens used. The dataset is tagged with synthetic, distilabel, and rlaif. The dataset is split into a training set with 500 examples.