thinQmaster Virtual testing partner Adaptive practice exams that find your weak spots and drill them. Every answer explained. From $69
Enroll

thinQmaster

Find your exam

Five questions free, before you decide.

Step 1 · choose a vendor

In one of our camps? Your practice exams come with the tuition — they are on your study page already.

AI & Machine Learning · Intermediate

Jailbreak

A technique that manipulates a model into bypassing its safety training to produce prohibited content.

Definition

Jailbreaks use role-play, hypothetical framing, obfuscation, many-shot examples or multi-turn escalation. Defenses include safety training, input and output classifiers, and continuous red teaming.

Exam tip

Jailbreak targets the model's safety policies; prompt injection targets the application's instructions.

Related terms