Since they're testing the sensor and not the final output of the camera, I'm pretty sure they are NOT using EXR mode. I think EXR mode would be considered in-camera processing. This particular series of tests doesn't concern itself with that - only the RAW output of the sensor. Still, it would be nice if DXOmark would make that clear. But unless you RAW exclusively, these tests don't seem all that relevent to me.
Brian, EXR Modes operates on sensor and are hardware methods to control DR or Higher ISO noise. Its not like Nikon or Sony's different in-camera processing for image enhancement. If it would be the case, no one should be complaining.
Truely, DXOMark scoring is wierd. Their numbers goes up and down too quickly. They are the only one who are testing sensor output, so no one is here to challenge it. Their methods are still kept secret (more or less) and not exaplined well, not like other review sites, those go through images in RAW and JPEGS and tries to prove their points. Also they delay their test results (for few manufacturers, specially fujifilm) so much, that mostly manufacturers by that times announced their next product line.
As a result, APSCs are scoring higher than Fullframe and higher end comapcts are scoring higher than big sensor compacts, if not overall then many occassions.