Fair enough, I should have been more precise: By "all possible tensor products" I actually meant all elements in the tensor product space, including all linear combinations of tensor products of single-particle states.