Something seems to be contradictory in this case.

1) If the model is a little nose heavy (manual specifies 3-1/8" behind LE at center line of the wing), nothing like that should happen.

2) If the model is way nose heavy, and is still slow after take-off, you may be forcing the take-off with low speed and big AOA.
When elevator is full up, with big deflection to rise the nose,
a) The tail could stall, inducing a nose down pitch that reduces AOA and downwash, re-attaching the airflow over the elevator, and then the process repeats.
b) May be the wing is starting a stall and falling forward, just to resume flight while the speed increases, and then the process repeats.
Sort of what happens to a chuck glider that has too much elevator up or aft CG.
The problem should improve while the speed increases, since both AOA's don't need to be high.

3) If the model is neutral or close to be tail heavy, and is still slow after take-off, it may become elevator sensitive or even unstable in pitch right after take-off,.............but it should continue behaving like that or worse while the speed increases.

I agree with da Rock and MinnFlyer, in the belief that #2 above is the case here.
However, for that to be truth, the specified CG must be incorrect.