Summary

The network supports black-box, white-box and transfer attacks, returning robustness scores, vulnerability heatmaps and hardening datasets to model owners. Perturb says findings will be privately disclosed and that the system is intended to provide continuous external testing rather than periodic red-team engagements.

Why it matters

A standing marketplace for adversarial testing could broaden access to model-security research, although its effectiveness and incentive quality remain to be demonstrated.

Sources