← Back to Glossary

LLM Jailbreak

A technique that circumvents the safety restrictions of large language models to produce outputs the model was designed to refuse.

Related Terms

AI-Powered Attacks Deepfake Adversarial Machine Learning Prompt Injection Data Poisoning Model Extraction AI Jailbreaking LLM Security Synthetic Identity Fraud Quantum Computing Threat 5G Security Blockchain Security