MoeReward/combined_preference_dataset_olmoe_base_math_heavy
收藏Hugging Face2025-03-21 更新2025-04-19 收录
下载链接：
https://hf-mirror.com/datasets/MoeReward/combined_preference_dataset_olmoe_base_math_heavy
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含两个部分：选中的(chosen)和被拒绝的(rejected)。每个部分都包含内容(content)和角色(role)两个字符串字段。整个数据集只有一个训练集(train split)，包含10000个示例。数据集的总大小为约15MB。
The dataset consists of two parts: chosen and rejected. Each part includes content and role string fields. The entire dataset has only one training set (train split) with 10,000 examples. The total size of the dataset is approximately 15MB.
提供机构：
MoeReward
