Two errors were discovered in our manuscript. Our chemical potential was shifted by 0.02t* for the half-filled case, resulting in a nonzero thermopower, when it must vanish by symmetry. Also, our formula for the electronic thermal conductivity was missing a factor of T-2 in the denominator.