Extension7/TinyWorld_Cities_8000
收藏Hugging Face2025-11-02 更新2025-11-15 收录
下载链接:
https://hf-mirror.com/datasets/Extension7/TinyWorld_Cities_8000
下载链接
链接失效反馈官方服务:
资源简介:
TinyWorld Cities 8K是一个包含8000个由GPT-4生成的简短城市描述的合成数据集,每个城市有五个主题的描述,总共约40000行。该数据集用于测试小型语言模型从合成文本中学习语法、结构和节奏的能力,并以JSON Lines格式存储。
TinyWorld Cities 8K is a synthetic dataset of 8,000 short city descriptions generated by GPT-4, with each city having five descriptions across different topics, totaling approximately 40,000 lines. This dataset is designed to test how small language models learn grammar, structure, and rhythm from synthetic text, and is stored in JSON Lines format.
提供机构:
Extension7



