MoeReward/combined_preference_dataset_qwen2.5_base_math_heavy
收藏Hugging Face2025-03-29 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/MoeReward/combined_preference_dataset_qwen2.5_base_math_heavy
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含两部分：选中的(chosen)和被拒绝的(rejected)内容，每个部分都有内容和角色两个字段。数据集仅包含训练集，共有10000个样本。
The dataset consists of two parts: chosen and rejected content, each part has content and role fields. The dataset only contains a training set with a total of 10,000 samples.
提供机构：
MoeReward
