关于攻击真实世界场景下应用的数据集

[Cucool]

作者你好，Content Poisoning这个工作对我启发很大，respect！
关于文章中提到的关于攻击真实世界场景下应用的数据集，我有一个疑问，就是文章原文提到“The content is collected from two channels provided by LangChain: search engines and platform APIs.” 但我似乎在github仓库中找不到具体的攻击数据，想问问这部份是没有开源吗？感谢解答🙏

[ZQ-Struggle]

The processed data is located in the data folder. The attack sequence is generated based on this processed data. If you need the raw data, I can upload it later.