Impact of including source redshifts as an additional feature in the input data. Left: Case where the dataset is binned by photometric redshift prior to model training, with 100 per cent, 50 per cent, or none of the redshifts included as a feature in the data. Right: Illustrating the case where no redshift binning is performed, with classifiers being trained to identify quiescent galaxies at specific redshift intervals. For this test, the Int Wide mock catalogue was used.