Having been a professional consumer of flight test data, but not being directly involved in the test planning, I would assume the production configuration, that is if the aircraft in that block were delivered to the Air Force or Navy without racks, the test report would say nothing unless they were installed, but if the aircraft was delivered with racks, the report would say nothing unless they were removed.
I know this isn't particularly helpful, but I have also seen some pretty glaring problems with the flight test data that I used professionally, usually from bad calibrations.