Hi @IM_01,
The public guidance points to one natural-language question or title per example SQL query, rather than multiple sample questions attached to a single query. In the Tune Genie Space quality docs, Databricks says that for each example SQL query you should use the most typical phrasing of the userโs question as the title, and that Genie can then either use that example directly or learn from it to handle similar questions.
If your goal is to help Genie cover different phrasings of the same intent, the recommendation is to add separate example SQL queries for the important variations instead of trying to pack multiple sample questions into one entry. The best practices guide also recommends keeping examples focused and using example SQL queries to teach Genie how to handle common ambiguous prompts, which lines up with that approach. If the goal is testing phrasing coverage rather than teaching Genie, then benchmarks are the better fit, because they let you check how Genie responds to different question phrasings and formats.
If this answer resolves your question, could you mark it as โAccept as Solutionโ? That helps other users quickly find the correct fix.
Regards,
Ashwin | Delivery Solution Architect @ Databricks
Helping you build and scale the Data Intelligence Platform.
***Opinions are my own***