BitAbuse
收藏资源简介:
BitAbuse数据集是由韩国中央大学等机构创建的,包含325580条句子,其中包含26591条视觉干扰文本句子和298989条非视觉干扰文本句子。该数据集是基于bitcoinabuse[.]com网站收集的与比特币欺诈相关的电子邮件构建的,旨在为防御网络钓鱼攻击提供真实世界的视觉干扰文本数据集。数据集通过人工标注的方式生成了每个字符的标签,以还原非干扰版本的文本。
The BitAbuse dataset was developed by institutions including Chung-Ang University in the Republic of Korea. It contains a total of 325,580 sentences, among which 26,591 are visually deceptive text sentences and 298,989 are non-visually deceptive text sentences. This dataset is constructed from Bitcoin fraud-related emails collected from the bitcoinabuse[.]com website, and is designed to provide a real-world visually deceptive text dataset for defending against phishing attacks. The dataset generates character-level labels through manual annotation to restore the original non-deceptive version of the text.




