Johannes Treutlein comments on Stop-gradients lead to fixed point predictions

[Johannes Treutlein]

You are right, thanks for the comment! Fixed it now.