Strengthening our defenses against distillation attacks
Illicit distillation allows model developers to copy frontier AI capabilities without creating proportionally robust safeguards. Here, we describe how we’ve reduced the rate of distillation by more than 90%.
Summary
Researchers have made significant progress in combating illicit distillation attacks that threaten AI model integrity, achieving a reduction rate of over 90%. Their findings highlight new strategies to protect advanced AI capabilities from being replicated without adequate security measures.
Comments (32 of ) ⌛