I think you're getting caught up in that transition time here. The wing lifts upward in a vertical line along the center line of the wing. If we roll the plane then the wing no longer lifts directly upwards. Instead the wing is now lifting to one side at the angle of bank. The resulting side component to the lift is going to accelerate the model sideways and thus we get our side slip and our dihedral induced righting force. Granted there will be a brief moment where the force has not yet accelerated the model but it's not long enough for anything to happen. Certainly the model doesn't snap back to level during THAT moment but rather after a 1/2 second or so when the side slip is well developed.

Other proof that there is a side slip happening when you bank like this is the fact that the nose of the model falls off into a spiral dive. This is the fin doing it's job as a result of the side slip. If there was no side slip happening then the nose wouldn't fall off into a spiral. Rather the model would just continue happily onwards with the wings tilted.

Take out fictitious model. You bank the wings with the ailerons then neutralize them. The model tilts then it starts to fall off into a diving turn (the first part of a spiral dive induced by the side slip), the model then levels the wings thanks to the side slip acting on the dihedral after some part of a diving turn has occured and the extra speed from the dive brings the nose up past level into a partial stall after which the model returns to stable flight with a oscillation or two. If there was no slide slip the model would not enter the spiral. Rather it would continue straight ahead.

Fair enough?