Even though both names are somewhat generic, not very descriptive, and technically correct, I think changing the figures' term “composition polynomial” to “transition polynomial” makes sense because
they are visually and semantically between the “transition constraints” and the “transition quotient(s)”, and
they are already called “transition polynomials” in the text (where, in fact, the term “composition polynomial” does not appear).
Even though both names are somewhat generic, not very descriptive, and technically correct, I think changing the figures' term “composition polynomial” to “transition polynomial” makes sense because