when we are calculating the differentiation of e^(-z) then its differentiation will be -e^(-z) then why not considered -ive sign
Why not take -ive differentiation
Actually there are two negative signs and that’s the reason they got compensated,
sigmoid function = (1+e^-z)^-1 and when you take its derivative applying chain rulr, one -ve sign comes for the negative power of (1+e^-z) and other negative sign comes from e^-z and hence they both results in net positive sign.
Hope that helped 