Not quite sure exactly how you mean, but one could add a small trim in series with one the psuedo CT arms to offset it a little, for balance reasons.
I'm not sure how useful that might be - there can be a little unbalance there for sure .... the signal phases probably won't exactly matched for voltage ... and gain of the ax7 for sure can be a little different .. and finally, the 'bal attenuatator' can be different again, depending on the type (potentiometers or switched resistors etc).
I like to use those inexpensive 2deck 24pos rotary switches from eby in this role .. you can get them already soldered up with good resistors like Dale for log or lin in some common values .. my own ones were a little less 'well assembled and the law was a little 'lumpy' too.
Well worth the investment in a solid unit ... otherwise dual potentiometers are mostly fine.
The upshot I think is that the balance is going the net of all that ... perhaps you can trim that up a bit .. certainly if you use dual-deck-potentiometers ... but it will vary for different attenuation levels.
And, after all that, the sidechain signal is mashed about a lot anyway ... with the dc bias and so on ...
Still, if I was going for precision as best as possible, I would consider it.
Good luck with it ... it's almost time for me to have another go at it
I'm still exploring mine [adaption of the fc660] , thru 'measurement' as well as in normal usage. I still can't get over how 'clean' [low distortion] the signal amp is, with the limiting action turned off.
Super low THD considering the strong levels flowing, in my collection of 4x frame grid pentodes in triode mode per phase, - near idle, its nearly a pair of 6L6s worth ... I think mine was 45mA per phase at the signal amp output traffo primary.