PhD candidate @ University of Oslo. Interested in AI safety and mechanistic interpretability. Working on formal proofs of representation properties in deep neural networks.