Phase-Altered Interleaved Randomized Benchmarking for Compiled Non-Clifford Gates
This research introduces Phase-Altered Interleaved Randomized Benchmarking (PA-IRB) to test if virtual non-Clifford phase gates affect gate error estimates. By comparing a phase-stripped implementation with a phase-dressed version of the same compiled gate, the study found that agreement in error rates suggests these virtual operations do not measurably increase the physical error budget under specific compilation stacks.