Online Safety Monitoring for LLMs
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 external model into an alarm decision by thresholding, with the threshold calibrated via risk control. In experiments on mathematical reasoning and red teaming datasets, we show that this simple design is competitive with more advanced monitors based on sequential hypothesis testing.
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 58%Verisight →
- PossiblePossibly related (embedding) · 47%IEEE Rolls Out Large Language Models Virtual Training Course →
- PossiblePossibly related (embedding) · 45%eval-harness-plus →
- LinkedLinked via arxiv author · 85%Mona Schirmer →
“Online Safety Monitoring for LLMs”
- LinkedLinked via arxiv author · 85%Metod Jazbec →
“Online Safety Monitoring for LLMs”
- LinkedLinked via arxiv author · 85%Alexander Timans →
“Online Safety Monitoring for LLMs”
- LinkedLinked via arxiv author · 85%Christian Naesseth →
“Online Safety Monitoring for LLMs”
- LinkedLinked via arxiv author · 85%Maja Waldron →
“Online Safety Monitoring for LLMs”