selfrew/llama3_test_pack_train_ratio035_epoch3_tmp0
收藏Hugging Face2024-12-16 更新2024-12-21 收录
下载链接：
https://hf-mirror.com/datasets/selfrew/llama3_test_pack_train_ratio035_epoch3_tmp0
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含5000个训练样本，每个样本包含多个字段，如问题（question）、真实推理过程（gt_cot）、真实答案（gt）、难度级别（level）、类型（type）、解决方案（solution）、奖励（rewards）和用户解决方案（my_solu）。这些字段表明数据集可能用于训练和评估问答系统或推理模型，特别是那些需要理解和生成复杂推理过程的模型。
This dataset contains 5000 training samples, each with multiple fields such as question, ground truth reasoning process (gt_cot), ground truth answer (gt), difficulty level (level), type, solution, rewards, and user solution (my_solu). These fields suggest that the dataset is likely used for training and evaluating question-answering systems or reasoning models, particularly those that require understanding and generating complex reasoning processes.
提供机构：
selfrew
