PPL
Sanmi Koyejo
Papers (6)
Is Backpropagation Optimal? When Synthetic Gradients Improve Sample EfficiencyThe Distillation Game: Adaptive Attacks & Efficient DefensesGeneral Preference Reinforcement LearningWhy Do Safety Guardrails Degrade Across Languages?Reasoning Models Don't Just Think Longer, They Move DifferentlyStop Automating Peer Review Without Rigorous Evaluation