Mark Barton created a Mathematica utility which pointed us in the direction of the UIM wire clamp locations on the mass as being a culprit in our asymmetric UIM mass.  I'll let him add the details of the cool model to the alog, but steal the nutshell for monolithic sake:

One thing that leapt to our attention is how sensitive the pitch is to the uimfixedfwdmm slider.  That means if the triple hang tooling bolts to the UIM in a slightly different place than the real clamps, there could be major discrepancies. 0.1 mm there is good for 14 mrad at the UIM, 7 mrad at the PUM and 5 mrad at the TM, which is rather more like what we're seeing than any prism or ear perturbation. 
-M. Barton

So, this afternoon we unclamped the triple hang tooling wire clamps and played with them in the mating grooves on the UIM.  We even switched the "right" segment for the "left" and then rehung the monolithic in the triple again.  We made no adjustments to the monolithic/UIM suspension itself.  This time, the pointing of the masses was slightly different than the morning measurements, indicating that in fact there is too large a range in something related to the triple hang tooling wire segments to get reproducible pointing.  (Note, the differentials clock together in magnitude as expected, they are in a different range than before.  From Jason's notes:

    UIM: ~1.739 mrad down
    PUM: ~0.505 mrad up
    ITM: ~2.677 mrad up

    UIM/PUM Differential: ~2.244 mrad
    PUM/ITM Differential: ~2.172 mrad

Note, we have not been able to reproduce the 5mRad differential UIM/PUM pitch noted last week, which has since gone awol.