遇见数据集

qfq/geminiall_hardfiltered_v2

收藏
Hugging Face2025-01-15 更新2025-04-12 收录
官方服务:

资源简介:

该数据集包含了多个字段,如解决方案(solution)、问题(question)、cot类型(cot_type)、来源类型(source_type)、元数据(metadata)、cot(其类型为null)、思考轨迹(thinking_trajectories,序列字符串)和尝试(attempt)。数据集分为训练集(train),共有24517个示例,大小为3755534082.2582307字节。配置信息中包含了默认配置和数据文件的路径。

The dataset contains multiple fields such as solution, question, cot_type, source_type, metadata, cot (with dtype as null), thinking_trajectories (sequence of strings), and attempt. The dataset is split into a training set (train) with a total of 24517 examples and a size of 3755534082.2582307 bytes. The configuration information includes the default configuration and the path to the data files.

提供机构:
qfq
二维码
社区交流群
二维码
科研交流群
商业服务