23 - Mechanistic Anomaly Detection with Mark Xu
AXRP - the AI X-risk Research Podcast
00:00
The Mechanistic Anomaly of Diamonds
AI is doing mechanistic anomaly detection to the composition of all three networks yes. But then I have this question of like okay I suppose in order to keep the diamond in the room they I has to come up with a new type of plan and there's a new threat that appears. So we just take the action our policy generates and we just put that in our predictor but we don't include the policy itself. Then we just ask does this action have the consequence of the diamond being on camera for the normal reasons diamonds sometimes appear on camera or not, maybe the action has paid someone else to smuggle in a fake diamond etc. We can try to point more precisely at what we mean by
Transcript
Play full episode
