I made myself a ZVex Wooly Mammoth clone, using the Madbean layout, and found the Bite control (equivalent to Pinch on the actual commercial product) to yield an interesting effect. It is pleasingly interactive with the Gain control setting.
This past week, I was dickering around with a clone of the JSH Fuzz (AKA Jen Fuzz III). It is essentially another silicon Fuzz Face, and uses a 150k resistor between the emitter of Q2 (actually Q3 in the circuit) and the base of Q1. As anticipated, increasing the feedback resistance yielded some nice changes to the intensity of the fuzz produced. I recommend such a mod to any Fuzz Face derivative. Does it need to be a pot? Not necessarily. For instance, you could stick a 470k in series with the stock 100k feedback resistor (or whatever value is used i that particular build), and use a 3-position toggle to select either the full 470k, bridge the 470k to revert back to stock, or stick another resistor in parallel with the 470k to yield something in between. Given how interactive it is with the Gain control, restricting yourself to "only" 3 positions does not limit the variety of sounds, so much as make it easier to get back to the ones you like.
Noting that the Dunlop Jimi Hendrix JH-1 Fuzz Face sticks a small-value cap in parallel with that resistance, it might even be worthwhile to tinker with use of a DPDT on-off-on toggle to stick different caps in parallel with that path to complement the resistance change.
Worth exploring.