aladinDJ/smoltalk-annotated
收藏Hugging Face2025-10-16 更新2025-10-25 收录
下载链接:
https://hf-mirror.com/datasets/aladinDJ/smoltalk-annotated
下载链接
链接失效反馈官方服务:
资源简介:
SmolTalk-Annotated是一个经过Magpie标记的扩展多轮对话数据集,基于原始的SmolTalk数据集构建,包含了详细的任务类别、输入质量、指令奖励、对话深度、安全性等多维度标注信息。这个数据集适用于数据质量分析和大型语言模型(LLM)的后期训练研究。
SmolTalk-Annotated is an extended multi-turn dialogue dataset that has been fully tagged with Magpie, built upon the original SmolTalk dataset. It includes detailed annotations for task category, input quality, instruction reward, conversation depth, safety, and more. This dataset is ideal for data-quality analysis and post-training research for large language models (LLM).
提供机构:
aladinDJ



