[Bug] Controlled operations are not right

[chMoussa]

If the order between the control is before the target for cases CNOT(control=0, target=1), everything will be correct. For cases such as CNOT(control=1, target=0), this will be incorrect as the matrix should be different.

Expected output of _controlled:
CNOT(control=0, target=1)
jnp.array([[1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 0, 1], [0, 0, 1, 0]])
and for CNOT(control=1, target=0):
jnp.array([[1, 0, 0, 0], [0, 0, 0, 1], [0, 0, 1, 0], [0, 1, 0, 0]])

Currently, for the last case, we obtain the same as CNOT(control=0, target=1):
jnp.array([[1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 0, 1], [0, 0, 1, 0]])