Fig. 2: GST (solid lines) and PAPA reconstruction (dashed-lines) Nσ vs. germ power L for each of the three experimental data sets.
From: Bootstrapping quantum process tomography via a perturbative ansatz

Nσ quantifies the deviation from a Markovian qubit model. The goodness-of-fit parameter is provided by pyGSTi and ranges from single blackstar to five black stars, indicating how well the data fits the implicit model. Larger values of L correspond to increased sensitivity to gate error and to longer circuits.