Projects

AI risk demo

This project aims to replicate the results from the Armstrong's toy model of reward hacking on LL...

BabelBack

Hyperthesis

Superposition

Sensemaker

Dialectic

AI TTX demo