diff --git a/tests/test_optimize.py b/tests/test_optimize.py
index 21b7a91..83523f9 100644
--- a/tests/test_optimize.py
+++ b/tests/test_optimize.py
@@ -8,14 +8,29 @@ def test_merit_function_view():
     def my_function(x):
         return x
 
-    x0 = [0., 0., 0.]
+    x0 = [0.0, 0.0, 0.0]
 
     opt = xd.Optimize.from_callable(my_function, x0=x0, steps=[1e-6, 1e-6, 1e-6],
                                     tar=[0., 0., 0.], tols=[1e-12, 1e-12, 1e-12],
                                     limits=[[-1, 2], [-1, 4], [-2, 2]])
-    opt.solve()
 
+    # Check that Jacobian doesn't introduce side effects
     mf = opt.get_merit_function()
+    xo.assert_allclose(mf.get_x(), x0, atol=1e-6, rtol=0)
+    xo.assert_allclose(opt.get_knob_values(), x0, atol=1e-6, rtol=0)
+
+    jac_init = mf.get_jacobian([0.5, 2, -1])
+    xo.assert_allclose(mf.get_x(), x0, atol=1e-6, rtol=0)
+    xo.assert_allclose(opt.get_knob_values(), x0, atol=1e-6, rtol=0)
+    xo.assert_allclose(jac_init, [[1, 0, 0], [0, 1, 0], [0, 0, 1]], atol=1e-6, rtol=0)
+
+    jac_init2 = mf.get_jacobian([0.5, 2, -1])
+    xo.assert_allclose(mf.get_x(), x0, atol=1e-6, rtol=0)
+    xo.assert_allclose(opt.get_knob_values(), x0, atol=1e-6, rtol=0)
+    xo.assert_allclose(jac_init2, [[1, 0, 0], [0, 1, 0], [0, 0, 1]], atol=1e-6, rtol=0)
+
+    opt.solve()
+
     jmf = mf.get_jacobian([0.5, 2, -1])
     xo.assert_allclose(jmf, [[1, 0, 0], [0, 1, 0], [
                        0, 0, 1]], atol=1e-6, rtol=0)
diff --git a/xdeps/optimize/optimize.py b/xdeps/optimize/optimize.py
index 3a5d097..53316db 100644
--- a/xdeps/optimize/optimize.py
+++ b/xdeps/optimize/optimize.py
@@ -391,6 +391,7 @@ def __call__(self, x=None, check_limits=None, return_scalar=None, zero_if_met=No
     def get_jacobian(self, x, f0=None):
         if hasattr(self, "_force_jacobian"):
             return self._force_jacobian
+        prev_x = self._get_x()
         x = np.array(x).copy()
         steps = self._knobs_to_x(self.steps_for_jacobian)
         assert len(x) == len(steps)
@@ -409,6 +410,7 @@ def get_jacobian(self, x, f0=None):
             x[ii] -= steps[ii]
 
         self._last_jac = jac
+        self._set_x(prev_x)
         return jac
 
     def _clip_to_max_steps(self, x_step):
@@ -1295,6 +1297,7 @@ def target_mismatch(self, ret=False, max_col_width=40):
     def get_knob_values(self, iteration=None):
         """
         Get the knob values at a given iteration.
+        If no iteration is given, return current knob values.
 
         Parameters
         ----------
@@ -1308,7 +1311,7 @@ def get_knob_values(self, iteration=None):
         """
 
         if iteration is None:
-            iteration = len(self._log["penalty"]) - 1
+            return self._err._extract_knob_values()
         out = dict()
         for ii, vv in enumerate(self.vary):
             out[vv.name] = self._log["knobs"][iteration][ii]