Do what Sluckey said first.
Do you have a way to take a picture of the chassis & post here? If so, do that while taking care to show wiring running near the Volume pots, and from the volume pots to pin 2 of the phase inverter.
Assuming there's not a leaky cap causing a bias issue (which is what I think Sluckey was going after), the wiring to the phase inverter may be picking up hum. The schematic shows 5MΩ pots, which is a quite high impedance and will easily pick up hum from nearby a.c. wiring. Rather than just say "shield the wire" I'd be interested to know if anything runs too close to the Volume pots themselves, or if that wire runs too close/too parallel to heater or other a.c. wiring.
By grounding the grid of the phase inverter and killing hum, you've shown the hum occurs before the grid; however, turning down the volumes doesn't nix the hum (because of the backwards arrangement to isolate the channels). And because of that backwards arrangement, the wiring from the Volumes to the phase inverter grid is always a high impedance and will readily pick up noise interference.
Or if you're in the tinkering mode, you can just try wiring the Volumes like a typical volume control, perhaps also adding 470kΩ resistors from the wiper to the phase inverter grid. That's the only thing left in-circuit which makes sense if grounding pin 2 stops the hum but grounding pin 7 of the phase inverter does not.
