In the GPTs best practices section of the API’s documentation, there is a tactic called “Use inner monologue or a sequence of queries to hide the model’s reasoning process”. One advice is to instruct the model to put parts of the output that are meant to be hidden from the user into a structured format that makes parsing them easy. In one of the examples, they instruct the model to enclose some of the output into triple quotes ; here’s the playground example provided by the documentation.
However, when I try it, the output doesn’t include any triple quote, and I need to remind the model in another user prompt to include it. Is it supposed to be the case ? Or is it something that would work better with gpt-4 ?
I would expect the examples taken directly from the documentation to work directly.
Thank you !