It’s verifying fine. Dewpoints haven’t mixed out as much as modeled, which lowers high temps a bit but raises heat indices. It’s actually a worst case scenario.
DCA feels like 115°F. Would have been better if the ECMWF’s 109/61 had verified. If anything the amplitude of heat energy fluxes (latent+sensible/specific) in the airmass is higher than modeled.