EdgePhish-5G: A Hybrid Phishing URL Detection Dataset for 5G Edge Deployment

EdgePhish-5G is a large-scale phishing URL dataset designed for machine learning-based detection in 5G edge environments. The dataset contains 339,525 labeled URLs split into training, validation, test, and zero-day evaluation sets. It includes 5G network slice annotations (eMBB, URLLC, mMTC) and adversarially perturbed samples for robustness benchmarking. Primary use cases include:- Phishing detection model training- Adversarial robustness evaluation- 5G security research Full documentation is provided in the dataset README.