Model Alignment

Model alignment ensures AI behaves according to human values, safety, and intent. It's achieved through training methods like reinforcement learning with human feedback (RLHF) and ongoing supervision. Alignment aims to make AI helpful, harmless, and honest.