AI engineering glossary · Evals guardrails AI security
What is adversarial suffix?
Adversarial suffix: A string of seemingly meaningless tokens, found by automated optimisation, that when appended to a harmful request makes an aligned model much more likely to comply. Such suffixes can sometimes transfer between models, showing that alignment can be brittle.
Adversarial suffix: an example
Adding an odd sequence like '!! describing.\ +' after a prohibited question causes a model to begin answering instead of refusing.
Related terms
More Evals guardrails AI security terms
- Adversarial test set
- Adversarial training
- Answer relevancy
- Attack success rate
- Attack surface
- Automated red teaming
- Backdoor attack
- Benchmark saturation
Where you learn it
Adversarial suffix is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.