A vector is just a line. In the case of a thrust vector, it is a line through the centerline or axis of the motor, where the thrust is assumed to be acting (ignoring the small perturbations and large deviations in the real world, caused by things like debris passing through the nozzle or deposits on the throat or nozzle). For most model rocket designs this is right up the centerline of the rocket. Sounds like that for yours, at least the thrust vector will be within the vertical center plane and will not be pointing left or right.
If you design the motor mount to point straight ahead, then the offset location of the mount (say, 1 inch) will be the same all the way forward through the length of the model. Let's assume for this example that the center of gravity is on the main longitudinal axis and not above it or below that axis. As the thrust vector passes the center of gravity, it will still be 1 inch away. So you will have a de-stabilizing torque (or moment, in engineering-talk) of (N lbs max thrust)x(1 inch) = N inch-pounds. The duration of this de-stabilizing or upsetting moment will be the burn time of the motor. After that, your model will cruise straight ahead as normal, stabilized by the fins. So the critical period is really the fraction of a second during peak motor thrust, when your model might still be on the launch rod?
So you could side-step much of the problem by using a pair of launch lugs, spaced up and down the length of your bird, to hang onto the launch rod/rail, together with a reasonably long launch rod. With your largest fins oriented perpendicular to the expected direction of your thrust-induced pitch-up, they should also help a great deal to dampen any pitching motion.
Didja get all that?