interpretability

Is AI Just a Fancy Autocomplete?

LLMs work through statistical relationships in high dimensional latent space. We can show that this information influences how a model…

3 weeks ago

Interpretability: What AI Actually Thinks (Then Fixing It)

How can we fix the bugs in AI models until we understand them? Learn how we'll all be debugging our…

7 months ago

The Tyranny of Chat Interfaces to LLMs

So much of our interaction with LLMs is via a singular chat box. It doesn't have to be that way.…

1 year ago