Two things to unpack here
1) Correct SBU/SUB/etc off-grid behavior. The problem here is that whatever inverter this was, it bridged grid and inverter output. That is not supposed to happen for off grid inverters. But it is not clear how well policed that is. In principle since 6000XP is listed as off grid inverter and has US company backing it, there is some stronger guarantee. However that bakes in an assumption about what exactly 1741 guarantees for off-grid inverters. If it only guarantees momentary parallel or better, then there are POCOs that will still beat you down if they find out. I have yet to see a thread here that lawyers the 1741 (IE down to sentence diagramming and quoting of the book) to the point of establishing that it guarantees full air-gapping between grid and inverter. Rather, it is just accepted that it is the case.
If you double convert, physics and your own hands of constructing the double conversion, is the proof. Instead of lawyering, disassembling the AIO to confirm, or using social infrastructure (buying listed from domestic supplier) to gain confidence.
2) Unwanted exports. Depends on what you mean.
Zero export when operating in parallel? Physically impossible to prevent without doing things like setting minimum import amount, and still taking a risk (though probably reduced if you do that, and the size of import needs to be modeled for your house).
Zero export from unexpected momentary parallel? See (1)
MultiPlus and Quattro can be used in both on-grid and off-grid cases, so you need to have the settings right, or use external disconnecting relay. In the latter case you may not have proper SBU capability.