If it's climbing inverted even after you took out that down-thrust (and I'll venture a guess that it is), then the elevator in upright "neutral" flight is trying to push the nose down, meaning the tail is heavy. Probably just slightly. Once inverted, the extra weight not only drops the tail, but the elevator trim pushes it down (when inverted, down is up, and up is expensive!) and exaggerates it. Move the CG forward. If you wish to test this theory first, trim your elevator so it climbs some in normal flight, and then invert it. You'll find it doesn't want to climb inverted, or at least, it doesn't want to climb as much as it did. If reducing the down-thrust leaves it flying level inverted, then don't change the CG!
Knife edge weirdness - remember, you have a big engine, a big prop, and a short-winged plane that, when on its knife edge, no longer has the wings holding the plane up, so the prop that's cranking away and torquing it right along through that knife-edge can go right ahead and turn the plane without those pesky wings interfering. The prop is swinging to the right, and torquing to the left. Wings that are creating lift and holding a plane up are at the same time neutralizing the prop torque. Wings aren't really being used during knife-edge. Hope this helps.
The one thing to remember is that the Pitts is designed to be a powerhouse flyer, and the VV40CC gives it what it likes, but it's a high-performance bipe, and even on a full-scale with a big engine, the prop will do its best to turn the plane, if given the chance to slip off its leash for a few seconds. The pilot knows that's the price you pay for the joy-joy flying.
Jim
P.S. You folks may have heard I've decided to retire from the hobby. My Pitts (identical to yours, EGG) is up for sale, same setup - Valley View 40CC.