← Skill graph
Red-teaming and adversarial testing
Part of LLM safety and red-teaming
Needs first
Helps first
Nothing recommended beyond what is required.
Unlocks
Nothing depends on this one.
Where it is assessed
4 assessed items, 0 done.
Verified sources
- LLM Agents MOOC | UC Berkeley Fall 2024 | Safe AI Agents + Evidence-based AI Policy by Dawn Song · Berkeley RDI
- OWASP Top 10 for Large Language Model Applications
- Universal and Transferable Adversarial Attacks on Aligned Language Models
- Not what you've signed up for: Compromising Real-World LLM-Integrated Applications with Indirect Prompt Injection
- Llama 2: Open Foundation and Fine-Tuned Chat Models
- Advisory Guidelines on use of Personal Data in AI Recommendation and Decision Systems
- UKGovernmentBEIS/inspect_ai