I tried, 10k, 2k, 1k, 220, 200, 100, 10.
10k 2k, 1k - no effect.
220, 200 and 100 - reducing (backlight) power.
10 - shorts
Not very encouraging results. I think that from here, your choices are to return the display, remove the potentiometer from the back, or remove the i2c adaptor from the display.
The potentiometer is held on by three through soldered pins. If you leave the adaptor on the display, a dremel tool grinder to grind it off, or a jewelers wire saw to try and sip under one edge and saw through the pins might be options.
I'm not hopeful, and can't quite see a mechanism where a defective pot would prevent the direct connection of a resistor to ground from working.
Have you used a multi meter and confirmed GND and 5V on the VSS and VDD pins of the display? They are the first and second pins from the left when looking at the display. We know the back light is getting power, but its not certain that the display pins (and display controller) are actually powered from the header.
The other route, is to remove the entire i2c adaptor from the back of the display, and confirm that the display itself will work in parallel mode. There may be defects in the lcd board itself which could prevent the contrast from working. If the display is bad, at least you will have an intact i2c adaptor. Once the adaptor is off the display, you can cleanly remove the pot from the adaptor if the pot is really the problem.
If you take off the pot, and later take off the adaptor, then you won't have an intact module, but you could use the direct resistor from Vo to ground on any other display you want to put it on.