The dataset consists of template-based phishing login pages impersonating 123 brands. These samples were tested to evaluate the robustness of the existing phishing detection system. The samples are capable of bypassing state-of-the-art techniques such as PhishIntention, PhishPedia, and EMD.
