green_leaf comments on OpenAI’s GPT-4 Safety Goals