In this workshop, participants will receive a playbook and learn to test applications for vulnerabilities related to Large Language Model (LLM) attacks and simulate adversarial attacks against AI models with a free license for the SPLX.ai platform.
Attendees will learn:
The 6 layers of AI defense
The nuances of testing nondeterministic systems,
The difference between prompt injection attacks and jail breaking attacks,
How to create probes (test cases) for bias, toxicity, alignment, and safety.
Participants are encouraged to bring their own laptops but it is not required.