To add to Andre's notes above: It's irrelevant that the KHP solution (which is chemically ambiguous, is it K2HPO4 or is it KH2PO4? -- assuming the former, and converting to KNa2PO4, but calculations would also work equally well for converting KH2PO4 to KNaHPO4) *initially* has fewer than 4.166 mmol present as HPO4(-2) ion. As you start adding the NaOH, and converting it eventually to K3PO4, any bits that were initially present as H2PO4(1-), or PO4(3-), or H3O(1+), or OH(1-), will all "come out in the wash". In short, titration is all about "theoretical, complete" reactions. If you want a corny analogy, it's like walking a dog around the block: you can take a dog from any house, walk once around the block, and return it to the same house. The distance "around the block" (amount of titrant required) doesn't change, regardless of the house (equilibrium-driven side conditions) position selected initially.
Now, that's different than asking about pH values in the solution, since the actual [H3O(1+)] level is affected by the various equilibrium reactions the salt ions nominally present may have undergone. (Or, which house did you just return that dog to?)
Hope this helps!
Robert C.
12/23/13