Alignment and Adversarial Robustness: Are More Human-Like Models More Secure?

Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.