AI in a vat: Fundamental limits of efficient world modelling for agent sandboxing and interpretability

Add code
Apr 06, 2025

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: