In the situation of supervised Understanding, the trainers played both sides: the person as well as the AI assistant. while in the reinforcement Finding out phase, human trainers very first rated responses the model https://chatgpt-openia.net/login