Crab as you know I (used to) fly EC225 in oil and gas support. In that role perhaps 95% of the flight time is at MCP doing 145-150 in the cruise, and up to 165 on descent. So taking into account climb time, perhaps 90% at 145+. Obviously the cause of a double engine failure in the cruise is hard to envisage, but since we spend the vast majority of time in the cruise and double engine failure is something we are required to train for, it seems sensible to take the cruise case - and for example I guess the PHI 76 was probably doing 145 when it hit the bird.
Before we had the simulator we used to do this in the aircraft. But we were wimps and simulated auto entry purely by lowering the lever - engine control switches were not retarded due to the risk is using real OEI power. So the "failure" was pilot flying initiated - pretty benign you might think.
Nearly everyone got it right, flaring the aircraft as the lever was lowered to maintain 1g+. But a few (mainly fairly junior IIRC) pilots just lowered the lever rapidly and all hell broke loose with things floating up at sub-zero g including the pilots, then they would react to that and pull the cyclic back a bit, the disc would bite and the Nr would wind up at a phenomenal rate. I recall on my TRE check having a trainee do this and I just caught the Nr with a massive pull as it reached the max transient power off Nr. Scary!
Regarding the bit about Nr continuing to decay at zero g, obviously the nature of our auto entry means I have never seen this, but I think the aerodynamics of it make the consequence clear.
So in summary, from fast cruise rapid entry into autorotation can either be a benign affair if leading with the cyclic, or absolute mayhem if the collective is lowered without any cyclic input.
Thinking back to my days on the AS332L where we did retard the throttles to practice autos, with the entry from say 120kts the effect was less critical, but still noticeable. Entry without aft cyclic was messy rather than mayhem.