10% would also include automatically reviewed tweets for misinformation, covid, and any other language filters they have in place.

Not all reviewed-for-moderation tweets would happen because of user reports.