CroCoSum

CroCoSum是由布朗大学创建的一个专注于跨语言代码切换摘要的数据集，包含超过24,000篇英文源文章和18,000篇人工编写的中文新闻摘要，其中超过92%的摘要包含代码切换短语。该数据集通过收集solidot.org上的技术新闻摘要构建，旨在研究代码切换现象在跨语言摘要中的应用，解决现有数据集依赖翻译而忽视自然语言表达的问题。

CroCoSum is a dataset dedicated to cross-lingual code-switching summarization, created by Brown University. It encompasses over 24,000 English source articles and 18,000 manually curated Chinese news summaries, with more than 92% of these summaries containing code-switching phrases. Developed by collecting technical news content from solidot.org, this dataset is designed to study the application of code-switching in cross-lingual summarization, aiming to address the issue that existing datasets rely on translation while neglecting naturally occurring linguistic expressions.