#452 – Dario Amodei: Anthropic CEO on Claude, AGI & the Future of AI & Humanity

17926.452 - 17944.864

So, you know, the Arabic feature, the Hebrew feature, the base 64 features were some examples that we studied in a lot of depth and really showed that they were what we thought they were. It turns out if you train a model twice as well and train two different models and do dictionary learning, you find analogous features in both of them. So that's fun. You find all kinds of different features.