Errata:

• In ue_feature.pdf, the last point "Scaling UE density continuum to estimate UE uncertainty will underestimate said uncertainty by ×2" is a too strong of a claim. Rather, let's just say that scaling of the UE density by a constant factor is may not be equivalent to changing the UE tune.
• In ue_feature.pdf, plots for data and embedding with label "jets with UE denisty > 1 MeV" were effectively produced with "UE density > 0" (due to a bug the cut value was rounded to an integer). This doesn't significantly affect the plots, though.