How to the handle the context situation if your GPT3 model get stuck in a loop for a certain context?

I have written some instructions for the GPT3 model to stay in the context of the ongoing conversation but it is following that order so deliberately that it is just not leaving that context. So, is there any better way to handle context in the conversation between the bot and the user?

above image conversation continuation

Welcome to the forum…

What model and settings are you using? What’s your prompt look like?

I’m using davinci latest model and sending some instructions along with the prompt.

What are the settings - temperature, frequency_penalty, etc.

What is the exact prompt?