Typically there is no relief in the steering gear system. I can only speak to Ford units as that is where I worked. But I'm pretty sure that's typical of most systems.
There is a small bleed-off at the pump, but it is no where near large enough to provide compensation for the heavy volume the pump is capable of.
Allow me to go into much greater theory and detail; this will take a bit but will help you understand.
I will discuss the typical fluid-pressure type steering, regardless if it's rack-and-pinion or ball-and-nut; does not matter here for this discussion. (I am excluding hydro-boost type units at this point, but they are similar).
First of all, think of the demand upon a typical steering pump. When does a power steering pump have to provide the greatest amount of pressure assist? At parking-lot, slow-manuever speeds. If you've ever driven a manual-steer car, this is incredibly obvious. So, a power steering pump has to provide it's greatest flow and pressue at slow engine speeds which are most often submissive to the slow parking speeds. No one rev's their engine to 5k rpm just to turn into the garage or out of the mall, right? Hence, because the pump is belt driven by the engine, it is slave to the engine rpm. So, slow engine speeds (in the parking mode) equate to low steering pump speeds. But the DEMAND of volume and pressure are greatest at that same time. It's an evil necessity; during low speeds, the pump is being turned slowly, but must match the greatest demand. And therefore, a power steering pump is sized (by volume and pressure capacity) to meet those demands. The pumps (especially in large cars and trucks) is 100% capable at near idle. Ironically, once the car/truck is moving at any decent speed (and subsequently the pump is turning with the engine at 2k or 3k rpm) every bit of pressure and volume is now excess capacity and waste; it's the nature of the beast. They key to understand is that a power-steering system is under it's greatest demand at the slowest of speeds, and therefore is fully capable at near idle. Everything past that is a waste.
OK - to be able to turn the wheels of the Lincoln Town Car or Ram 3500 truck, there has to be full capacity available when the wheel is turned. But what of all that pressure and volume when the wheel is NOT being turned? It's dumped via hydraulic equalization (often called the center circuit) in the steering valve. The steering valves have a "t-bar" (torsion bar) that connects the steering input (a mechanical machined tube with lands and grooves) inside a "sleeve" (a mating part to the input with lands and grooves) while being seated inside a pinion or sector reciever. (You can google the picture for the details). Essentially the t-bar is pinned to both the input and the driver, with the sleeve floating around it.
When the wheel is not being turned, the fluid flows equally through the left/right circuits in the valve, relieving all that high-capacity volume and pressure. However, when you turn the wheel, the t-bar flexes, and allows the input to turn in relation to the sleeve, and closes off one of the two directive paths and fully opens the other one. (Think of it this way; imagine water flowing over a knife blade where the blade is equally splitting the water stream. Once you turn the knife one way or the other, the water will no longer be "split" and instead flow only in one direction, at the expense of the other. In the steering valve, the more you turn, the more the flow is taken from one side, and delivered to the other.) You have to think of the TOTAL flow in relation to the input and the sleeve. (I'll make up some numbers here; these are just bogus but they help with the concept). Think of 100 psi going into the input/sleeve, with 6 gpm entering, and 3 gpm exiting per side. Got it? When you "turn", it flexes the t-bar, diverts some differential to one side, and steals from the other. That now creates a pressure differential, that is sent to the steering gear (be it R&P or Ball/nut). That pressure acts against a piston, which pushes the steering (rack or sector) and turns the wheels.
When you turn the wheel, you typically only turn it some fractional amount of full-lock. If the steering system is capable of 6 total turns left-to-rigth, it might be easily split into 3 left and 3 right turns. So when you go around a corner, you may only use 1/2 rotation of the steering wheel, 1.3 turn, etc. But you typically don't "fuly lock" the steering wheel all the way over.
When you turn the wheel, the effort goes down the column, into the steering valve, and the pressure assists in turning the wheels, but that stops as soon as you stop putting pressure on the wheel. Even if you turn left for 50% of total rotation, the system reaches a "balance" because once you stop turning the wheel, the t-bar eventually evens out the hydraulic circuit. So the wheels stop moving left or right because you quit creating a differential in the input/sleeve circuit (you are no longer flexing the t-bar). The pump will only strain a very short time and then the wheels move to balance out the pressure differential. The system natually hunts out (seeks) a balanced state because the t-bar is trying it's best to even out the flow and "center" the wheels to the hydraulic and mechanical input. Even if you drove around in a circle for three hours, the input is satisfied as long as you don't increase or decrease the input differential. Only when you change the steering wheel position, does the valve have to react to the shift. (To be honest, there is a bit more complication than that, and the components are VERY tight in tolerances down to tenths of thoutandths, but you get the idea, right?).
And remember, the pump is capacitized for full demand, heavy load times.
All is well and good. You steer, the input/sleeve/t-bar react. Until ....
You turn the wheel all the way over 'til the steering linkage hits the steering stops. PANIC TIME! You are now creating a condition where the steering valve cannot equalize the pressure paths, because you're continuing to turn into a postive stop position. The valve cannot balance. Therefore, the pump dead-heads against the flow, and you nearly instantly cook the steering fluid.
But why no pressure relief valve? Now here is where my long diatribe gets to your meaty question (sorry it takes so long) ...
Here's why:
What happens in an emergency when you yank the wheel in avoidance of running into that car that pulled out into your lane, or that child that runs into the street? When you turn the wheel in such a violent manner, you absolutely need the system to react quickly, with no delay. When you turn so abrutly, you need 100% flow and volume to cause immedate reaction in the system. And, don't forget that at "normal" driving speeds, your pump is way over any needed capacity; it is essentially supplying WAY more volume/flow than is needed. So, the system cannot have a "bypass relief valve" in it; or it would relieve at the very time you cannot afford to have it do so. If you had a bypass valve that would safely dump off pressure at idle speeds, it would cause a non-steer condition during an emergency. If there were a bypass valve that was effective at slow speeds, it would cause the system to dump when you need it most, and the system is grossly over-capacitized anyway.
There were a few lame attempts to create some switched valve systems that would deactivate bypass only during emergencies, but those never caught on, and they never were accepted by the government FMVSS committees. (Federal Motor Vehicle Safety Standards). At least not that I'm aware of.
And so, you have a hydraulic system that has two evil conditions. You need it to be fully capable at slow speeds, but it can self-destruct if turned into a positive stop position, and it must be 100% absolutely functional during an emergency, when the system is WAY over capacitized; therefore no relief valve is accpetable to avoid a "loss of asssit" condition.
You still with me, or did I bore you to death?
On your way home tonight (legal disclaimer here; only do this when safe to do so), experiment a bit. In the parking lot, both quickly and slowly turn the wheel and listen to the pump response. And, then, ONLY VERY BRIEFLY - less than 1 second turn it to full lock and hear the pump labor. You can feel the hydraulic power that is present. Then, get up to speed and violently turn the wheel a few times. Feel how much excess is in the system at normal driving speeds speeds. Play with the system a bit, and then consider what I've explained, and it all makes sense. No one would truly want the system to relieve itself at the very moment you need to it function the most.
