I do not think this is a perfect way to look at it, but it is my first cut.
To solve the lack of data about class (FR, SO, JR, SR) I narrowed my focus to the four time AA (excluding 2020). For Myles Amine I only used his first four years.
There have been 154 four time AAs. But treating them as a group when looking at paths is problematic because the guys who started by winning their freshman year cannot improve upon that performance. So I segmented the group by starting points (freshman finish).
Obviously, the guys who won it all as a freshman perform the worst on average their sophomore year as their delta can only be zero or negative.
A couple observations:
Freshman AAs, who go on to AA four times, are pretty evenly distributed among the places.
That qualifier in the previous sentence is doing a lot of work. Missing from this analysis are guys who AAed as freshman, but did not AA all four years. That puts a potentially significant upward bias on all the deltas above.
Surprising to me, on average second through eighth place finishers as freshman improve on that placement as sophomores.
Broadly speaking, in this limited data set, yes, AAs do tend to improve in their career. But again, this is just true for four timers.
For giggles, here are the four timers who won it as freshmen: