I do not think you will find torque specs for the saddle. Screw sizes and mechanisms differ significantly for different types of seat post clamps.
Arky is right. Asking for a "torque spec" for a saddle is nonsensical; the saddle doesn't have any threaded parts on it.
Obviously, you're wondering how much you can crank on your seatpost clamp bolt(s) without crushing the carbon rails. Frankly, I wouldn't worry about it. I've never heard of anyone crushing carbon saddle rails before (breaking, yes, but not crushing in a clamp). Besides, the carbon rails are either very thick-walled or solid; either condition would make it harder to crush the carbon than, say, a thin-walled titanium tube. Just tighten your seatpost clamp until it doesn't slip and be done with it.
Lots of consumers take specified torque as gospel, but it's much less exact than some people think. The concept of "maximum torque" is really a stand-in for maximum clamping force. The relationship between applied torque and resulting clamping force depends on lots of things, especially these three: Mechanism design, thread pitch and lubrication (or the lack thereof). I'll discuss these in more detail:
- Mechanism design: Depending on how the saddle clamp is designed, you can generate lots of clamping force or relatively little from a given amount of torque. For example, a Campagnolo Record seatpost uses a single bolt, and the clamping force has essentially a 1:1 ratio with the force applied by tightening the bolt. On the other hand, the USE Alien post has a unique clamp design (they call it the "cyclops" clamp) that generates a lot of clamping force with just a little torque. I'm just guessing here, but it might be a ratio of 3:1 or 4:1. If I recall correctly, the Campy post has an M8 bolt while the USE post has an M5 bolt. Even if I got the bolt sizes wrong, you get the idea: the Campy post has a much beefier bolt because it needs a relatively high amount of torque to clamp tightly; the USE post has a much smaller bolt because it generates a similar clamping force with less applied torque.
- Thread pitch: A finer thread will generate a lot more clamping force than a coarser one. Many seatposts use M6 bolts; these have a thread pitch of one thread per millimeter (10 thread peaks or valleys per cm); M8 bolts have either 1 or 1.25 mm between each thread (8 or 10 threads/cm); M5 bolts have a 0.8mm thread pitch (12.5 threads/cm); finally, M4 bolts have a 0.7mm thread pitch (~14 threads/cm). The finer the thread pitch, the more clamping force you get for a given torque. My Ritchey WCS stem uses M4-threaded bolts and specifies a max applied torque of 5 N-m. That generates a higher clamping force than my old 3T stem which used M5 bolts and also specified a max torque of 5 N-m.
- Lubrication: This sounds obvious, but lots of people miss it: a lubed bolt tightens more easily than a dry bolt. If you apply 5 N-m to a dry bolt, it will stop turning when the friction force generated by the bolt's threads turning against the female threads creates a reaction torque of 5 N-m. If you lube the threads, the friction doesn't ramp up as fast as it did with the dry bolt, and so you get more clamping force for the same amount of applied torque. If you lube the threads but not the underside of the bolt head, you will get significantly less clamping force than you would have if you lubed up all the contact surfaces. A dry bolt head against a dry surface will generate lots of friction and stop your bolt from turning.
It's safe to assume that manufacturers' torque specs are for fully lubricated bolts and threads. (It doesn't take much grease to make a bolt "fully lubricated.")
As you can see, a torque spec takes into account lots of different variables. While an engineer can control the clamping mechanism and the thread pitch, the lubrication aspect is fairly variable. Some of my engineering work has involved aircraft control systems; these are parts in which improperly tightened bolts could cause a loss of control and a crash. In order to make sure I'm getting a certain clamping force in a bolted joint, I'll sometimes specify a number of turns after contact. If I have a clamp closed by an M6x1mm bolt, I might want that clamp to close by 3.5mm in order to be fully tightened. In this case, I would specify 3.5 turns after contact. (The bolt has a 1mm thread pitch, so each turn moves the bolt by 1mm).
Even the "number of turns" (NoT is the acronym I just made up) approach is imperfect...if my clamp is designed right but the part it's clamping is slightly undersized, the specified NoT won't produce enough clamping force. Tightening the bolt to a specified torque would produce more clamping force on the undersized part (whether it's "enough" is a separate question) than tightening to a specified NoT. Because a stem manufacturer can't assume what brand and model of handlebar might be bolted into his stem, torque specs make more sense than NoT specs.
You can make this as complicated as you like: e.g., bolted joints with are more flexible with Ti bolts than they are with steel bolts, which can affect how much clamping force is developed.
The point is that torque specs are not purely objective. Rather, they're approximations that include lots of assumptions. That said, I'm glad that more manufacturers are making them available.