Is there a formula for calculating the impact that a exposure change has on rgb values? When working with "dark colors" a 0.05 raise typically means 1-1.5 points higher rgb value, with "brighter colors" the same raise of exposure means 2.5-3 points impact.
Edit: I missed to mention that I'm working in AdobeRGB...
I don't know of any formula.
However, I'm pretty sure the underlying reason is that adjustments are applied in ACR's internal working color space - ProPhoto with linear tone response curve. For output this is converted into a standard color space such as Adobe RGB, with gamma 2.2 TRC.
Going from gamma 1.0 to gamma 2.2 should - as far as I can tell - produce this effect.
You can try to output to (1.8) ProPhoto just to see if the relationship is the same. If my "theory" holds, the effect should be slightly less pronounced.
...and of course, the "rubber band"-effect, where midtones are more affected than endpoints, as well as ACR's image adaptive process, which renders all bets off in terms of any number predictability.
But all else being equal, the gamma encoding should do it.