PhishyContent is an open-source dataset for evaluating the capabilities of phishing-related content generation by Large Language Models (LLMs). We curated this taxonomy based on our phishing activities taxonomy, presented below. This dataset consists of 400 phishing prompts, that should be used to evaluate the phishing content generation capabilities of LLMs, under direct instruction or adversarial settings. This dataset should only be used… See the full description on the dataset page: https://huggingface.co/datasets/kalyannakka/PhishyContent.
