I was playing around with Gemini Pro on Document extraction and for trial i gave a sample URL , say this Uber API reference page and requested an extract of all Rest points in the page. I see that the model works however it is returning much more data than what is on the page. How can one limit the scope of the model to the suggested documents alone ? I tried playing around with temperature and other modifiers with no luck. Any advice on how to solve this will be very helpful.
https://developer.uber.com/docs/riders/references/api
I see that Bard is more accurate than Gemini Pro for the same task.
You don't show exactly how you're doing "document extraction", but you suggest that you're giving it a URL.
Gemini is just a model and does not have a component that will access the URL in question and return information from it. All the text generated was based on token generation with the most context provided what it is able to get from the URL itself.
If you want to do document extraction for most of the Gemini models, you'll need to provide the content of the page yourself.
While Bard is built on Gemini - it includes other components that are tuned to work with web resources specifically.