If there's a blank row, or a question cell seems to contain something that isn't a question (like a section header) as judged up front by the large language model, then the system should just ignore that row and move on. Right now it stops right at the blank or header row and does not proceed with questions below it.
If there's a blank row, or a question cell seems to contain something that isn't a question (like a section header) as judged up front by the large language model, then the system should just ignore that row and move on. Right now it stops right at the blank or header row and does not proceed with questions below it.