eigvals!(A) vs eigvals!(A, B)?

What is the difference between eigvals!(A) and eigvals!(A,B)?

I understand that eigvals!(A) will replace A with a “scaled” version, where rows and columns are more equal in norm and the matrix becomes closer to upper triangular (and maybe other stuff). This is documented.

I did not find eigvals!(A,B) in the documentation. What is B used for?

Similar to eigvals(A,B) and eig(A,B), it solves the generalized eigenproblem Ax=λBx.