Language model evaluation

1. Defensive Distillation: AI (Brace For These Hidden GPT Dangers)
2. Generative Teaching Networks: AI (Brace For These Hidden GPT Dangers)