Inside the Hidden Geometry of AI

Goodfire is an AI interpretability research company. We had a nice paper a few months ago called reinforcement learning with feature rewards, where we were able to use the internal concepts of the model that we were able to extract in order to reward the model to reduce hallucinations.

It goes beyond the title with direct discussion of like, model, models, including: So we spend all day thinking about what's going on inside the mind of a neural network.

1. 01We had a nice paper a few months ago called reinforcement learning with feature rewards, where we were able to use the internal concepts of the model that we were able to extract i
2. 02It's the idea that models really think in shapes and not really in words
3. 03You can use this to help debug and train your model to be more effective and safe and performant