Mapping the Mind of a Large Language Model Anthropic

Interesting analysis from Anthropic figuring out how to model out the "feature" components of an LLMs data (brain?) and then modify the priority of a "feature" to change the way in which it engages.

The fact that manipulating these features causes corresponding changes to behavior validates that they aren't just correlated with the presence of concepts in input text, but also causally shape the model's behavior. In other words, the features are likely to be a faithful part of how the model internally represents the world, and how it uses these representations in its behavior.

Someone call up a neuroscientist for AI!