I can't provide references specific to this manoeuvre. The main texts on helicopter dynamics only generalise since such a manoeuvre isn't flown in real life or tested during certifications.
When you raise the collective, you increase the pitch of the all of the blades together. In most flight regimes, this means increased angle of attack and therefore lift.
The cost of producing lift is drag which always increases. So, there is a direct relationship between the collective lever and power required from the engine.
As you continue to raise the collective, AoA and lift will increase until you reach a point where the power required to overcome drag exceeds the rated power of the engine. Most helicopters permit operation beyond the maximum for a short time, e.g. 2 or 5 minutes, typically during take-off. Continue to increase AoA beyond this limit and drag will exceed power available and the rotor RPM will drop (a reduction in rotor RPM caused by drag exceeding power through too much collective is known as "droop").
As RRPM decreases, so does lift and the helicopter will descend. The relative airflow into the disk now has a greater horizontal component and the angle of attack will increase, either causing yet more drag or stalling the rotor.
So there is some point, before the collective is pulled fully up where maximum lift is attained. Beyond that, lift reduces until recovery or a stall. The only time I have ever raised the collective fully is during "full and free" movement checks with the engine off as part of pre-flight.
Your hypothetical manoeuvre:
From VNE in level flight, you roll into a 70 degree bank.
The TRT( total rotor thrust) is always acting perpendicular to the plane of the rotor disk. Since you are in a bank, and I assume it is at constant altitude, there is a large horizontal component of the TRT to maintain the acceleration in the turn and a large vertical component to oppose gravity and maintain altitude.
This requires a lot of power which means a lot of collective and a lot of drag. I don't know for the AH-6, maybe there's a chart relating bank angle and speed to power but let's just say, your collective is going to be somewhere near the max lift point with a manoeuvre like this.
Since you are now accelerating with max power, your speed will decrease. Therefore lift will decrease and since you are at or near max power, you will begin to descend.
So far, this is probably recoverable simply by rolling level and reducing collective until back inside the performance envelope.
The second part is where it definitely will go wrong. In all helicopters with a non-rigid head, including the AH-6, the disk will adopt a new plane of rotation rapidly. There is then a significant delay before the fuselage adopts a new attitude which puts the TRT perpendicular with the aircraft Z axis.
Pulling hard on the cyclic will very rapidly put the disk into a new plane of rotation which has a large angle between the plane of rotation and the fuselage. Additionally, the disk loading will increase dramatically and the blades will be forced downwards. On an articulated head such as that on the AH-6, droop stops will prevent the blades from bending down so much that they fold.
Hitting the droop stops a little is uncomfortable with a lot of vibration. Pulling hard on the cyclic at VNE in a 70 degree bank at or near max power may cause failure of the droop stops or blades or cause the static droop stops to contact the mast possibly leading to rotor separation.
Let's assume that there is no failure. The plane of rotation will be at the maximum permitted by the physical limits of the flapping hinges which is greater than the angle between the plane of rotation and the fuselage.
Therefore, I would expect a hard pull on the cyclic to result in a blade hitting the tail or the top of the cockpit. Either results in destruction of the aircraft, either as it breaks up at that point or when it hits the ground 10 seconds later.