nicher92/magpie_llama70b_60k_filtered_swedish_DPO
收藏Hugging Face2025-02-13 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/nicher92/magpie_llama70b_60k_filtered_swedish_DPO
下载链接
链接失效反馈官方服务：
资源简介：
这是一个经过质量筛选的Magpie数据集子集，每个输入生成5个答案并通过奖励模型进行排名，包含输入和响应的生成配置信息，以及任务类别和难度等字段。数据集分为训练集，共有约61568个示例。
This is a subset of the Magpie datasets filtered by quality, generating 5 answers per input and ranking them using a reward model. It includes input and response generation configurations, as well as fields for task categories and difficulty. The dataset is split into a training set with approximately 61,568 examples.
提供机构：
nicher92
