mlfoundations-dev/REASONING_evalchemy

该数据集包含两个配置：default和instances。每个配置都包含文本内容（content）、角色（role）、生成参数（gen_kwargs）、重复索引（repeat_idx）、请求索引（request_idx）、任务名称（task_name）和元数据（metadata）。元数据中包含期望答案、问题ID和参考解决方案等。数据集仅包含训练集split，default配置有3127个示例，instances配置也有3127个示例。数据集的下载大小和总大小分别给出。

The dataset consists of two configurations: default and instances. Each configuration includes text content (content), role (role), generation parameters (gen_kwargs), repeat index (repeat_idx), request index (request_idx), task name (task_name), and metadata. The metadata contains expected answer, problem ID, and reference solution, etc. The dataset only contains a training set split, with 3127 examples in the default configuration and 3127 examples in the instances configuration. The download size and the total size of the dataset are provided.