We have imaged GM Aur with HST, detected its disk in scattered light at 1400A and 1650A, and compared these with observations at 3300A, 5550A, 1.1 microns, and 1.6 microns. The scattered light increases at shorter wavelengths. The radial surface brightness profile at 3300A shows no evidence of the 24AU radius cavity that has been previously observed in sub-mm observations. Comparison with dust grain opacity models indicates the surface of the entire disk is populated with sub-micron grains. We have compiled an SED from 0.1 microns to 1 mm, and used it to constrain a model of the star+disk system that includes the sub-mm cavity using the Monte Carlo Radiative Transfer code by Barbara Whitney. The best-fit model image indicates that the cavity should be detectable in the F330W bandpass if the cavity has been cleared of both large and small dust grains, but we do not detect it. The lack of an observed cavity can be explained by the presence of sub-microns grains interior to the sub-mm cavity wall. We suggest one explanation for this which could be due to a planet of mass <9 Jupiter masses interior to 24 AU. A unique cylindrical structure is detected in the FUV data from the Advanced Camera for Surveys/Solar Blind Channel. It is aligned along the system semi-minor axis, but does not resemble an accretion-driven jet. The structure is limb-brightened and extends 190 +/- 35 AU above the disk midplane. The inner radius of the limb-brightening is 40 +/- 10 AU, just beyond the sub-millimeter cavity wall.