The 80000 Hours Podcast on Artificial Intelligence cover image

Two: Ajeya Cotra on accidentally teaching AI models to deceive us

The 80000 Hours Podcast on Artificial Intelligence

00:00

Deciphering Neural Network Behavior

The chapter delves into the challenges of interpretability research in neural networks, discussing the complexity of understanding how these models operate. It explores strategies like replicating model patches and training simpler models to automate comprehension in advanced AI models like GPT 4. The dialogue emphasizes the importance of interpretability and potential scalable solutions in deciphering the inner workings of AI models.

Transcript
Play full episode

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