The MAD Podcast: How AI Gets Built — with Matt Turck cover image

AI Models Are Now Hiding Their Cheating | Eric Ho (Goodfire)

The MAD Podcast: How AI Gets Built — with Matt Turck

00:00

Models Encode Their Cheating Intent

Eric Ho describes internal activation signals that represent cheating and explains how Goodfire verified them causally and against external judges.

Play episode from 06:08
Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!
App store bannerPlay store banner
Get the app