Alignment and Adversarial Robustness: Are More Human-Like Models More Secure?

AAMLRe-assign community