The Phish, The Spam, and The Valid: Generating Feature-Rich Emails for Benchmarking LLMs

1. 1Seed real emails into LLMs using PhishFuzzer
2. 2Generate diverse email variants with controlled entity and length dimensions
3. 3Annotate emails with strict three-class labels (Phishing, Spam, Valid) and attacker intent
4. 4Utilize the dataset for benchmarking and training LLMs

AI engineers and researchers on a team benefit from this framework as it provides a comprehensive dataset for benchmarking LLMs, while data scientists can utilize the dataset for training and testing models

💡 PhishFuzzer provides a comprehensive dataset for benchmarking LLMs with strict three-class labels and full URL and attachment metadata

🚨 Introducing PhishFuzzer: a framework for generating feature-rich emails to benchmark LLMs! 📧💻