PhreshPhish is a large-scale, real-world dataset and benchmark for phishing webpage detection containing phishing and benign HTML-URL pairs.
Train 371,941 samples: 253,189 benign and 118,752 phish
Test 36,787 samples: 30,048 benign and 6,739 phish
Benchmarks 404 benchmarks with baserates ranging from [5e-4, 1e-3, 5e-3, 1e-2, 5e-2]
v1.0.0 (2025-05-14): Initial release
from datasets import load_dataset
train =… See the full description on the dataset page:
https://huggingface.co/datasets/Yadhu-Vipin/phreshphish.
