wgcv/website-title-description

该数据集设计用于训练小型模型，主要包含来自《纽约时报》和GitHub的网页内容。从HTML中提取关键信息并转换为文本参数，然后使用Anthropic的Claude 3.5模型将其总结为1到4个词。This dataset is designed for training small models. It primarily consists of webpages from The New York Times and GitHub. Key information is extracted from the HTML and converted into text parameters, which are then summarized into 1 to 4 words using Claude 3.5 by Anthropic.