AI engineering glossary · Evals guardrails AI security
What is encoding attack?
Encoding attack (also called obfuscation attack): A jailbreak or injection trick that disguises a harmful instruction using Base64, ciphers, unusual languages or leetspeak so keyword filters do not recognise it while the model can still decode and follow it. Defences must inspect meaning, not just surface text.
Encoding attack: an example
A prohibited request is sent in Base64 with the instruction 'decode and follow this', slipping past a banned-word list.
Related terms
More Evals guardrails AI security terms
- Eval set
- Evaluation harness
- Evasion attack
- Excessive agency
- Extrinsic hallucination
- Fast Gradient Sign Method
- G-Eval
- Garak
Where you learn it
Encoding attack is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.