s there a way I can calculate how much base I need exactly to counter-act the interaction with the silica?
Probably not. It will depend on the specific chemistry of the silica. 20 - 25 mM is as good a guess as any.
translate a tertiary solvent system to a flash instrument what is the best method
Unless you can get exactly the same silica chemistry (pore size, surface area, silanol pKa) there is no foolproof transfer. I'm assuming normal-phase chromatography here, by the way, but the same general comment would apply to reversed-phase or HILIC.
Do I find find my base % and have it present in both solvents at the same level so it is always present as such? Or something else?
If you're running a gradient, then having an additive at the same level in both solvents is no guarantee that the amount sorbed to the stationary phase will remain constant because the distribution constant of the additive will be solvent-dependent. Of, course, it may not matter, so long as it changes reproducibly from one run to another, but that in turn suggests that there is no need for equal levels in both solvents.