Besides, nothing else is modified to correct for this tile-warping effect. Deep-space SBs have a radius of dozens of light years, while starbases next to stars have a radius of only a light year or so. Seems a little hard to square that when dealing with the gravity well projector etc.
Which starbase effects need modification?
Culture starbase effects? Influence doesn't make any real sense anyways, so who cares.
Economic starbase effects? The only sensible way for a starbase factory or resort to function is for there to be cargo/passenger ship traffic between the starbase and the worlds under its effects; increases in absolute tile dimensions are cancelled out by the corresponding increases in ship speed. Labs need to be sufficiently close that you have at minimum a sufficiently low-latency communications link between the starbase labs and the planetary labs to enable effective collaboration, though you likely have that to most if not all points within your empire anyways; the more sensible limiter on lab range is once again the need for ships to go to and from the starbase, this time carrying test samples and other resources for the labs instead of raw materials for the factories and refineries (and refined or finished products back to the planets), and so you once again have a situation where the variation in absolute tile dimension is cancelled by the variation in ship speed. Markets can either go with the communications link or with actual ship traffic; once again, if it's ship traffic you have a situation where the variations in actual tile dimensions are cancelled by the variations in ship speeds. I don't see a sensible way for a trade route value bonus to work except if the station works as a handling facility, which means ship traffic. Nothing here is difficult to square with the effects of mass upon tile dimensions.
Mining stations? The only rational way for a mining station to function is to send out mining ships to the resources within range and have them gather resources and haul them back, or alternatively for it to function as a handling point which both provides the supplies for and collects the resources from smaller mining outposts at the actual resource site. The variations in speed and tile dimension cancel out, so there's no issues to resolve here.
The modules for investigating relics and ascension crystals? Probably the same deal as mining stations; you're either sending probes and low-duration expeditions to or supplying research outposts at the relics, in which case it's ship traffic, or you're establishing an on-station lab which analyzes some kind of sensor readings (if so, then likely with special-purpose sensors rather than the ones used for detecting ships; after all, you can use the station to study a relic without it being within 'normal' sensor range).
Military starbases? The effects of military starbases are not sensible, at least not by the implied causes to which the effects are attributed (i.e. station-based effect generators). You used 'inertial dampeners' or 'mass-lightening technology' to allow your mass drivers to be accelerated to much greater velocities than they should be capable of moving at? Great, but when you let that mass come back, guess what? You (should, if it were to make sense) lose the extra speed, which means you may as well have just used a lighter projectile in the first place. No, the attack and defense boosts offered by starbases make far more sense as a "we can expend munitions and resources more freely because we can be resupplied more rapidly from the nearby starbase" than anything else, and resupply depends on the ability of some kind of cargo craft to traverse the distance between the station and the ship(s) under its effects. If you do want to go with the implied causes of the station bonuses, then the justification for the station bonuses can be summed up as "magic," and as far as I'm aware "magic" does not need to follow real-world logic, obey any process which can be explained by science, or operate according to the limitations imposed by reality.
As far as the speed modifiers of a military starbase, I would ask where adding a large amount of mass (or adding something which has the same apparent effect as adding a large amount of mass) makes a bigger difference - in a region with a lot of mass already present, or in a region where very little mass is present? I would expect it to matter most where little mass is present. If you model the warping effect as a right triangle with x being the distance between two points in 'real' space, d being the apparent difference in elevation in naturally-warped space as seen by the hyperdrive, and h being the increase in apparent difference in elevation in artificially-warped space (i.e. space affected by a Gravity Field Generator on a Military Starbase), where does a constant h have the greatest effect upon the apparent distance between the two points as seen by the hyperdrive? Why, where d is small, i.e. where the 'surface' of space is 'flat' - in other words, in places where there isn't already a large amount of mass present. At larger values of x, the station's effects are simply being obscured by the preexisting effects of the stars and planets already in the region. If space is a sheet of rubber and a gravity field generator is a marble, which changes the sheet's deformation more? Adding 1 marble to an existing collection of marbles on the sheet, or adding 1 marble to a sheet which does not have any marbles upon it?