I'm looking at the source code for the Qiskit Experiments module and see that the EFRabi class subclasses the Rabi class with the intent of calibrating rotations between the 1 and 2 states instead of between 0 and 1. The only change I see in the circuit is the addition of an X gate at the beginning. How could this be done for other calibration experiments, e.g. one of the FineAmplitude experiments? Why does the X gate allow us to calibrate on higher energy states?