Fetching the paper…

Adversarial Robustness Limits via Scaling-Law and Human-Alignment Studies · Around