Atticus Geiger

Researcher who aims to explain artificial neural networks by uncovering interpretable symbolic algorithms they implement, using formal models of causality and a theory of causal abstraction to understand the relationship between neural networks and high-level algorithms, with work intended to make models more reliable, safe, and trustworthy.

No endorsements yet.

Personal Website

https://atticusg.github.io/

LinkedIn

-

Twitter / X

-

LessWrong

-

EA Forum

-

No grants recorded.