Low frequency is not a neutral property of a letter. It changes how the letter behaves, and the iotified big yus is the clearest case in this alphabet.
Frequent letters are stable. A scribe writes them thousands of times, sees them thousands of times, and any drift is corrected by the sheer weight of examples. Rare letters have none of that. They are written from memory of a rule rather than from habit, they are the ones a tired copyist substitutes for something similar, and the substitutions are not caught because the reader has no strong expectation either.
The result is a compounding problem for anyone studying such a letter. The evidence is thin, and it is also worse in quality than the evidence for common letters, because the errors are concentrated exactly where the practice is weakest. Statements about the shape, the distribution or even the existence of a rare sign in a given tradition therefore rest on a handful of attestations that may themselves be mistakes.
The practical consequence for this alphabet is that the four yus letters are ordered by how much can be confidently said about them, and the two iotified ones come last. The atlas records their derivations as debated for that reason as much as for any argument about Glagolitic: the evidence for the shapes is not good enough to support a firm claim in any direction.
Saying so is better than picking a reading, because a rare letter with a confident history is usually a rare letter somebody has overinterpreted.