Please explain as best you can how to get the answer(with theory) and I will give you a thumbs up. Don't simply use chat GPT.

b) A recurrent neural network (RNN) has an input weight matrix U, a recurrent weight matrix W, and an output weight matrix V. Given an input vector xt at time step t, the output of the RNN at time step t is given by ot=Vst where st=f(Uxt+Wst1), and f() is a nonlinear activation function. Given a vector sequence X={x1,,xT} the output oT can be obtained by unfolding the RNN for T time steps. When T is large, it is more appropriate to use sigmoid or hyperbolic tangent instead of ReLU as the nonlinear activation function. Briefly explain the reason. (6 marks)