Deflection of rudders can balance moment about the centre of gravity. I have thought of non-symmetric deflection of rudders so that net sideward force is create to balance the moment about the center of gravity of the aircraft.
The Hornet has a fly-by-wire flight control system that is run by 2 flight control computers (also known as Roll Pitch Yaw computers) and these each have 2 channels. This system gets pilot input from the controls but the computers will decide the most efficient way to impart the command signal using all of the available control surfaces. Pilot has no say in which surface moves. If they did, this aircraft would be a nightmare to control because, like most modern fighter jets, it has been designed to be inherently unstable. This allows it to be highly manoeuvrable during a dogfight.
Some of the surfaces are controlled by 2 channels (rudder, aileron, leading edge flaps). Some are quad redundant, utilising all 4 channels (Trailing Edge Flaps, Horizontal Stabs).
All of the surfaces, with the exception of the leading edge flaps from memory (because they are connected mechanically with a rod???), can deflect independently of each other, if the software/flight envelope permits. For example, if you want to roll a Hornet, some combination of all or some of the following surfaces may deflect: L and R Aileron, L and R Trailing Edge Flaps, L and R Horizontal Stabilisers. There are various sensors such as accelerometers which provide the negative feedback loop required to ensure the aircraft is getting to the required attitude at the expected rate. If it isn't, surfaces may deflect more or others may be commanded to help out. The circuitry inside the computer is basically a summing amp with the following inputs: command signal from the pilot, the position feedback from the surface and the rate feedback from the accelerometers. I think the important thing to understand is that this signal is nulled (0 volts) when the aircraft is pointing in the commanded attitude. That means no more movement of the surface.
With regards to rudder on takeoff, these accelerometers would sense an uncommanded change in the yaw axis and I suspect the flight control computers would damp this affect using the feedback method described above without the pilot really noticing much, certainly not having to stomp on a rudder pedal to compensate. There would be some sort of signal processing delay and mechanical reactive delay to correct the movement but this is all done without pilot input.
The rudders don't have to deflect symmetrically as they are controlled by different flight control computer channel pairs, one from each computer for redundancy. In fact, when the rudder fails, it is commanded to streamline and the other surfaces will compensate. The software programming inside the flight control computers that determines which surface/surfaces move and by how much.