When I type in a prompt using our chatGPT plugin and then when I log the queries that the plugin is receiving the queiries look like simplified versions of what I'm prompting with. For example, if I type 12 words, the query will be simplified to a 4-5 word sentance. So they are extracting only what "it believes" is the key info. But I am wondering if we are able to make it use the full the initial message to generate a complete response based on our exact prompt, without omitting details. Can we override this, so that we get an exact answer to our exact question, not a simplified version that is interpreted by ChatGPT. In our case quantity is very important, let's say we need a quantity of 10, chat gpt will ignore this and send a quantity of 4-5. How can we make it understand that quantity needs to be exact?
For example here is my prompt:
Can you recommend me a Mexican Restaurant near Toronto?
And here is the ChatGPT response:
GET /openai.yaml HTTP/1.1" 200 OK
PROMPT: Mexican restaurant near Toronto
Any help here is appreciated to get more accurate results. Thank you!
