beccabai/slimpajama_labeled
收藏Hugging Face2024-10-21 更新2024-12-14 收录
下载链接：
https://hf-mirror.com/datasets/beccabai/slimpajama_labeled
下载链接
链接失效反馈官方服务：
资源简介：
该数据集用于论文《Multi-Agent Collaborative Data Selection for Efficient LLM Pretraining》，是SlimPajama-627B训练数据集的标注版本。数据集包含文本生成任务，语言为英语。每个数据项包括ID、内容和元数据，元数据中包含了属性如教育相关性和主题分类。数据集还定义了不同主题的标签，如活动、教育、娱乐等。
This dataset is a labeled version used in the paper Multi-Agent Collaborative Data Selection for Efficient LLM Pretraining, based on the SlimPajama-627B train dataset. It is designed for text generation tasks in English. Each sample includes a unique id, content text, and metadata, which includes attribute scores and domain information. Additionally, the dataset provides a mapping of topics and labels, covering categories such as activity, education, entertainment, finance, health, business and industrial, infrastructure, literature and art, nature, others, law and government, networking, and technology.
提供机构：
beccabai
