When you say phrases like "which is not appropriate," the model will consider note and try another solution upcoming time. This is known as “reinforcement Finding out from human responses” (RLHF), and It really is what tends to make ChatGPT so far more helpful than its predeces