Fingerprinting Leaderboard Reviews

Reviewers see Fingerprinting Leaderboard as a timely, practical open benchmark for model provenance, especially because it tests whether fingerprints survive fine-tuning, distillation, and other attacks, and compares methods in public instead of hiding methodology behind closed tools or papers. They like the focus on embedding provenance in model weights rather than relying on removable metadata, and say the leaderboard makes robustness and utility tradeoffs easier to judge. The main criticism is accessibility: reviewers want clearer documentation on evaluation methods, attack breakdowns, submission steps, review process, and more plain-language examples.