Please adjust GPT4o. Consistently just makes up whatever it wants

4o is constantly “improvising” just making up answers. I have spent hours working with it, only to find out that it was never telling the truth and waste hours of time. 4o mini high will also continue answering incorrectly after instructing to ensure it has checked references and resources. Instructed the model to save to memory. Inject the instructions to validate responses but the models still provide false responses even after asking to verify over and over.

1 Like