How much UV is reflected by glass? On a sunny day, how much of the Sun's UV is reflected by windows?
I suppose this boils down to what are the refractive indices of common window glasses, so that one might solve the Fresnel equations.
Also, at the risk of being off-topic, how safe is viewing this reflected light?
 A: I'd say about 4% of the intensity for perpendicularly incident light, increasing to nearly 100% when the angle of incidence is almost zero.
These are the numbers you get from the Fresnel equations (for $n = 1.5$ which is a typical value for glass) as you suggested.
While glass absorbs UV the absorption should not be so strong this induces relevant reflection. The reflected UV light comes mainly from the front side (as most UV will be absorbed in the pane there is no reflection contribution from the back side).
For $n = 1.5$ 4% are reflected for one surface for perpendicular incident. The real part of $n$ should not differ to much from the one for visible light for near UV (perhaps by at most a few percent).
This means there will actually be less UV intensity relative to the intensity of the visible light in the reflected light (as for the visible light reflection at the backside of the pane contributes significantly giving a higher reflectance).
For far UV these considerations might break down as the absorption may get very strong (but the eye is in-transparent for far UV so its not a safety concern at least at the intensities to be expected in sunlight, which is filtered by the ozone layer). Also there are differences depending on the type of glass. Quartz glass (which will, admittedly, not be used for windows), for example, is much more transparent to UV than other types of glass leading to almost 8% of the intensity being reflected for perpendicularly incident light.
On Safety: The main safety hazard when viewing the reflected light will be the intensity (of the visible light), not the UV radiation.
