← The frontier
Technology & AIJul 2, 2026

Online Safety Monitoring for LLMs

Despite alignment training, LLMs remain prone to generating unsafe outputs at deployment time.

Despite alignment training, LLMs remain prone to generating unsafe outputs at deployment time. Monitoring outputs online and raising an alarm when safety can no longer be assumed is therefore critical. We study a simple real-time monitor that turns a verifier signal from an…

The frontier is open to all. Sign in to learn this from first principles and save it to your knowledge base.