It's not really valid to treat this as a sequence, since most side lengths don't form what we think of as a Sudoku puzzle. The side length of a true Sudoku must be a perfect square in order to create the interior boxes that are (geometrical) squares. For the counterexample, think of a 7x7 puzzle (or any prime)... how can a square box contain seven cells?
A 4x4 puzzle gives you four 2x2 boxes. The standard 9x9 puzzle gives nine 3x3 boxes. A 16x16 puzzle gives sixteen 4x4 boxes. Sudoku variations sometimes have rectangular boxes; a 6x6 puzzle can have six 2x3 boxes, or a 12x12 puzzle can have twelve 3x4 boxes. But that is a different form of constraint logic so we shouldn't expect that the minimum number of clues for these sizes would follow a recognizable sequence.
I had to check Wikipedia (which, as we all know, always is right :-)) to see that you are right. I have seen so many variations on sudoku's that I forgot what the original looked like. I was just thinking of Latin squares.
(tongue in cheek) right, a 1x1 grid requires 0 clues, it has only a single digit as its solution; that digit is -- [its_so_on looks left, cut to SatvikBeri]
For example, the fact that it is not very hard to proof that a 1x1 sudoku requires 1 clue does not prove that sudoku is not NP-complete.