PromptSiege — learn AI security by attacking a real on-device LLM.
Mathematical proofs and formal methods for ensuring AI system safety.
Understanding when to use testing versus formal verification for AI safety.
Using formal methods to prove safety properties of AI systems.