AXRP - the AI X-risk Research Podcast cover image

23 - Mechanistic Anomaly Detection with Mark Xu

AXRP - the AI X-risk Research Podcast

CHAPTER

The Problem With Mechanistic Anomaly Detection

The main risk is we can't talk coherently about like this for the normal reason for any reason thing at all because it might just not make any sense but I think yeah so I'm gonna sketch out a way in which I think this might make sense okay. We have two sensors sensing our diamond or human one of them is like an image sensor and one of them are like a weight sensor to see if they're correlated on a specific data point. So there's this interesting fact on the training distribution which is that these sensors are often correlated oftentimes they'll both say yes or suppose that 50% chance there's a diamond during training. And so naively you would expect their conjunction to

00:00
Transcript
Play full episode

Remember Everything You Learn from Podcasts

Save insights instantly, chat with episodes, and build lasting knowledge - all powered by AI.
App store bannerPlay store banner