Score Centering Stabilizes Off-policy Reinforcement Learning Paper • 2609.20807 • Published Sep 17 • 8
Running on Zero Agents 4 Detector Scope 🔬 4 Auditable AI-text detection with sparse-autoencoder features