Once the response is received from the openAI how to differentiate between the response which is from the context or if it is from the public knowledge ?