6 - Debate and Imitative Generalization with Beth Barnes
AXRP - the AI X-risk Research Podcast
00:00
How to Train a Machine Learning Model for Debate?
So for debate, is the idea that, like, i'm going to have some question that I'm going to ask a machine learning model. And two different models are gong to debate each other, and they're gong to read the transcript, and then i'mgoing to know the answer is that roughly right? That's only the aveal idea. So in particular, we probably want two copies of the same model. If you just did this once, you wouldn't have any particular reason to believe that the answers that you go wr right? Like, the idea is just that this providesa the correct training signal,. Such as if you train for winning debates for a
Transcript
Play full episode
