Hopfield Neural net cannot store more than 2 association patterns

I'm just trying to implement a simple discrete Hopfield net using matlab but there is a problem with my result. As I present two patterns to be stored in my net everything works great and I can retrieve the correct association pattern given a noisy input, but if I just decide to store or train more than 2 patterns to my net, by presenting any noisy input,the net converges to a nonsense result, I searched through the net and I noticed it may be because of "Spurious States" and "Spin Glass" but there is no tip or guidance for dealing with this issue.

I would be thankful to guide me in this case, here is my taring code based on Hebb learning rule (presented in the book written by Fausett)

function trainBtn_Callback(hObject, ~, handles)

   weightMatrix = zeros(100,100);
   for k = 1:size(handles.associations,1)          %train the network for each association
      currentA = handles.associations(k,:);       %specify the associationa pattern which
      for i = 1:100
         for j=1:100
            if i ~= j
                w(i,j) = currentA(i).*currentA(j);  %learning the weight matrix by Hebb
            end
         end
      end
      for i=1:100
         w(i,i) = 0;                    
      end
      weightMatrix = weightMatrix + w;            % sum of the calculated weight matrices
   end
handles.w = weightMatrix;
guidata(hObject,handles);
    % code
  end