Generally Intelligent cover image

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

CHAPTER

The Dialog Model Is Used to Predict What All the Players Are Going to Do

The dialogue model takes into account the like board state, the history, the history of dialogue in general that has happened across all the players. We feed that into a large language model and then we use that to predict what all the players are going to do for the current term. It also gives us the intents that we feed into the dialogue model. So it gives us the action that we're going to play and then theaction that we would like the other player to play as well. And this dialogue model is trained on the human data. We actually filter out deceptive messages from the training data for this dialogue generation part.

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.
App store bannerPlay store banner