A screenshot is aways helpful.
I think you are talking about an issue that has been rectified, although I am not sure it it has add it into the latest studio build.
The original way modifier wall count differences were applied was to use the modifiers shape as part of the change. It essentially p, carved the modifier out of the overlapping portion of the model forming walls where the edges met.
This would mean a square appearing in the model if that was the overlap.
The new algorithm version, does what you would expect it simply changes the number of walls within the overlapping parts.
This change doesn’t appear to be a setting though, this is disappointing as the issue you detailed allowed for some advanced patterns to be generated which subsequent to the fix will not be possible.
Example of a “feature” that benefits from the original method.
I would like to see a toggle that the user can choose which technique is used.