Adversarial examples and attacks

1. Perfect AI Alignment vs Imperfect AI Alignment (Prompt Engineering Secrets)
2. Inherent AI Alignment vs Learned AI Alignment (Prompt Engineering Secrets)