How to compute derivative of a neural network model at a particular point using MATLAB?

Question

In using neural networks for solving differential equations, using MATLAB, I could find derivative of the model using the command 'dlgradient'. i.e.

gradientsU = dlgradient(sum(U,'all'),{dlX},'EnableHigherDerivatives',true)

where U is the model found using fullyconnect and dlX is the input.

Now, how we can calculate derivative at a particular point of the model U. To be specific I wanted to add derivative of the model at a particular point, say U_0=U'(5) to the loss function. So, how can I compute that?

I have followed the documentation given in MATLAB, given by,

"To evaluate Rosenbrock's function and its gradient at the point [–1,2], create a dlarray of the point and then call dlfeval on the function handle @rosenbrock.

x0 = dlarray([-1,2]);
[fval,gradval] = dlfeval(@rosenbrock,x0);

But I have already called using dlfeval to evaluate the model, so, not able to call dlfeval again. And when I'm trying to directly compute using the command dlfeval(@U,U_0), the output is always zero.

It will be really helpful if some insight could be provided. Thanks in advance.