macabdul9/gpqa

这是一个包含多个特征字段的NLP数据集，其中包括数据源、提示内容与角色、能力、奖励模型（包括真实值、token数和风格）、额外信息（包括索引和分割信息）、响应序列、token长度和问题。数据集有一个测试集分割，提供了字节数和示例数。数据集支持默认配置。

This is an NLP dataset with multiple feature fields, including data source, prompt content and role, ability, reward model (including ground truth, number of tokens, and style), extra information (including index and split information), response sequence, token lengths, and question. The dataset has a test split with provided byte size and number of examples. The dataset supports a default configuration.