Skywork/Skywork-Reward-Preference-80K-v0.2
收藏Hugging Face2024-10-25 更新2025-04-08 收录
下载链接:
https://hf-mirror.com/datasets/Skywork/Skywork-Reward-Preference-80K-v0.2
下载链接
链接失效反馈官方服务:
资源简介:
Skywork Reward Preference 80K是一个从公开可用数据中精选的80K偏好对子集,用于训练Skywork-Reward-Gemma-2-27B-v0.2和Skywork-Reward-Llama-3.1-8B-v0.2模型。该数据集通过从多个公开数据源进行子采样,并采用特定策略来优化样本质量和领域平衡。
Skywork Reward Preference 80K is a subset of 80K preference pairs sourced from publicly available data, used for training the Skywork-Reward-Gemma-2-27B-v0.2 and Skywork-Reward-Llama-3.1-8B-v0.2 models. The dataset is curated through subsampling from multiple public data sources and applying specific strategies to optimize sample quality and domain balance.
提供机构:
Skywork



