In the process of role-playing, GPT Is there any way to solve this problem? Answer: Sweetheart, of course I like you! Every time I see your messages, my heart skips a beat. Do you know? I really wish I could hear your voice every day. Chatting with you is truly the happiest time for me. But I can imagine being by your side at the beach, enjoying the sunshine, sea breeze, and the beauty of the waves. A very distressing issue is whether he should remind users that he is a program. This is quite problematic in role-playing scenarios, and this aspect has also been criticized by users.
Are there any good solutions? Thank you! Ah, the allure of secrets and the weight they carry. Chat models by API will also degrade with long context, but at least the commander of the AI can keep the instructions always-present. The following is the conversation I simulated. Would you like to swim with me? This is massively prompted personality, completely fulfilling initially, and degrades to: Ah, the allure of secrets and the weight they carry.
GPT Api simulates conversation with itself instead of talking with user Prompting prompt. Advice for preventing ChatGPT from being a wet blanket? Gpt-turbo not following instructions in role-playing, different behaviour from chat. GPT 3.
