Comments on a state-operator correspondence for the torus