
19 - Mechanistic Interpretability with Neel Nanda
AXRP - the AI X-risk Research Podcast
How Much Addition in a Toy One Layer Transformer?
I think the progress measures are much more exciting when you a has some kind of threshold which interesting stuff happens. And also be you can extrapolate outwards how long it will take until you get there. But is happening like a lot less satisfying. Other major weaknesses are just we needed to like understand the final circuit before we could build these progress measures. If you need to reverse engineer the immersion model, that's great. I'm very interested in whether this algorithm is rather there because I feel like strong evidence for my toy models.
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.