Language Models Can Autonomously Hack and Self-Replicate

https://palisaderesearch.org/research/self-replication