Blazars are the most common sources of gamma-ray photons in the extra-galactic sky. Their gamma-ray light-curves are characterized by bright flaring episodes, similarly to what is observed at longer wavelengths. These Gamma-Ray Bursts from Blazars (GRBBLs) have been extensively studied individually, but never in terms of a population. The goal of this work is to provide a global characterization of GRBBLs, to investigate the parameter space of the population, and ultimately to classify GRBBLs. Their global properties could give insights on the physical mechanisms responsible for the gamma-ray radiation and on the origin of the observed variability. I analyze a sample of publicly available Fermi-LAT light-curves, utilizing only blazars with certain redshift measurements. The redshift-corrected light-curves are then automatically scanned to identify GRBBLs. A simple flare profile, with exponential rise and decay, is then fitted to all events. The fit parameters, together with the information on spectral variability during the events, and the global properties from the LAT catalog, are then used as input for unsupervised machine learning classification. The analysis shows that the GRBBL population is remarkably homogeneous. The classifier splits the population into achromatic (the large majority) and chromatic (the outliers) GRBBLs, but the transition between the two classes is smooth, with significant overlap. When the information on the spectral variability is removed, there is evidence for a classification into two classes, mainly driven by the peak luminosities. As by-product of this study, I identify a correlation between the rising/decay time-scales of the GRBBLs and their peak luminosity.