As the Wikipedia article says in the very sentence under the equations: if you are contracting over the wrong indices, you'll indeed pick up a minus sign. In that case, you are no longer using the Ricci tensor, because that's not the Ricci tensor anymore; in other words, you also draw the conclusion that Yang is wrong. (S)he mistakenly contracts over the wrong indices, and is thus a minus sign off. This is trivially proven false. G^uv=8paiGT^uv and G^uv=-8paiGT^uv can only both be true is G^uv=0, which obviously is not the case in general. You are wrong.