I know G changed some signs of some IEs a few times. Besides G model, Reinin's positivist and negativist dichotomy also built according to IEs of base signs. I think it perfectly correlates with model G. I am only not sure about the signs of Si, it may be different for SXI types and their supervision, dual's supervison chain. Sign of IEs may be the second important aspect of socionics after position and effects of position of IEs in terms of explaining the differences in types. Ofcourse every socionist can make up their own model, we can make ours right now in a way that signs can be different or irrelevant. However, I think the classical or modern socionics models (Augusta, Reinin and Gulenko) got it right.