mlfoundations-dev/GPQADiamond_evalchemy

该数据集包含两个配置：default和instances。每个配置都有context、gen_kwargs、repeat_idx、request_idx、task_name等字段。context字段包含content和role两个子字段，gen_kwargs包含do_sample、max_new_tokens、seed和temperature等子字段。在default配置中，metadata字段为null类型，而在instances配置中，metadata包含request_id字段。数据集划分为训练集，其中default配置的训练集包含594个示例，大小为499995字节；instances配置的训练集也包含594个示例，大小为491085字节。

The dataset consists of two configurations: default and instances. Each configuration includes fields such as context, gen_kwargs, repeat_idx, request_idx, and task_name. The context field contains sub-fields content and role, while gen_kwargs includes sub-fields do_sample, max_new_tokens, seed, and temperature. In the default configuration, the metadata field is of type null, whereas in the instances configuration, metadata includes the request_id field. The dataset is split into a training set, with the default configurations training set containing 594 examples and being 499995 bytes in size; the instances configurations training set also contains 594 examples and is 491085 bytes in size.