The short answer to your question is that you are running a finite element model that is not set up to calculate a global factor of safety. You should evaluate stability using this model in terms of convergence and predicted strain which is the purpose of finite element modeling. If you want a factor of safety, do a limiting equilibrium model. This is the best way to compute a factor of safety for a particular slope (Fs=resisting force/driving force).
The method that ntschwanz mentioned (strength reduction method) gets you a factor of safety, but this is not something that you can compare to a factor of safety obtained from a limiting equilibrium analysis. Generally speaking, when discussing factor of safety for a slope it's, by industry standard, determined using a limiting equilibrium model.