Episode 27: Noam Brown, FAIR, on achieving human-level performance in poker and Diplomacy, and the power of spending compute at inference time
Generally Intelligent
00:00
The Power of Large Language Models in the Planning Process
We use two architectures. One, we use like a large language model. This gives us the initial prediction. And we also use this for dialogue generation. We use a different architecture for the planning value function. If you were to use that in the planning process, it would be like 200 times slower. So the custom architecture allows us to do this planning much faster than what otherwise would possible.
Transcript
Play full episode
