With Kairos, the attack success rate on AgentDojo is 2.7%, not 47.0%
Kairos is a monitor that holds an effect that is not clear. It holds the effect until the next steps show if it is safe. The agent keeps its utility.
Sources
Posted
What matters in AI.
SubscribeKairos is a monitor that holds an effect that is not clear. It holds the effect until the next steps show if it is safe. The agent keeps its utility.
Posted
