Error tolerance in Newton Solver

Dear all,
is it possible to reduce the error tolerance in the Newton Solver which Dynare uses to simulate deterministic models? Many Thanks!
Best,
Ansgar