Understaindg the Assistant process

Does the assistant ‘learn’ from its mistakes. eg I have an assistant that has a word document uploaded and referenced in the Instructions. I asked it for some information about person ‘xx yy’ and it stated it did not have any information on that person. I informed it that there is information about that person in the document. It replied sorry for the in site. I asked again and it replied.

  1. Will this be remembered in other sessions.
  2. More importantly is asked about person ‘tt yy’ will it reference the document.
1 Like

The models do not learn from past mistakes unless you include the correction or new instruction in subsequent prompting.

2 Likes