Gaseous,
I am not suggesting aggressively pulling back. My advice was aimed more at what to do if you've got the collective fully down and your NR is decaying still--not dangerously low 'yet', but getting there.
However, I would still advise against pushing forward cyclic when in an auto if you're worried about low NR. Realize that the outer portion of the blades are wind milling, and providing the thrust (ie, their local total force vector has a forward (positive torque - driving NR higher) component in the disk axis). The magnitude of this force is a function of airflow through the disk, and thus reduction of flow through the disk as will occur if the disk is less perpendicular to the airflow will reduce this tip thrust and NR will decrease. Without this force the form drag on the blades will cause an NR reduction.
I understand your point about blade stall, but I don't think it should drive you to push forward for airspeed. The impact of reducing the airflow will have an immediate effect on the blade thrust and NR, but will have a slower response on airspeed. If you can get there, put it in a turn. Keep the nose down and airspeed up, but load up the disk with slight aft cyclic. You'll see a higher number on your VSI, but you'll have kinetic energy in your airspeed and in the rotor.
This is all based on my engineering background and flight experience. I don't mean to present this as fact. However, if I found myself in an auto with the NR going down through 80% I think I would still pull back gently on the cyclic. I'd like to look deeper into the numbers and perhaps model/simulate it computationally. My gut feeling is that in forward flight greater airflow up through the disk on the retreating side caused by pulling back on the stick would be better than the reverse flow already there and exacerbated by pushing forward cyclic. This is where I'm guessing stall would first show up, as opposed to the advancing side.
Don't try this at home
