Web Archive Singapore WARC Files
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/masamune-prog/warc2summary
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了112个WARC文件，这些文件经过GPT-4o处理，以生成元数据，重点在于降低成本和提高元数据生成的效率。此外，该数据集展示了大型语言模型（LLMs）在生成元数据方面的有效性，同时突显了内容不准确和虚构等挑战。规模上，数据集包含了112个文件，任务是对网络档案进行元数据生成。
This dataset comprises 112 WARC files, which were processed using GPT-4o to generate metadata, with a primary focus on cost reduction and improved efficiency in metadata generation. Additionally, this dataset demonstrates the effectiveness of Large Language Models (LLMs) in metadata generation, while also highlighting critical challenges including content inaccuracy and fabricated content. In terms of scale, the dataset contains these 112 files, and its core task is metadata generation for web archives.
提供机构：
Web Archive Singapore
