Variations in the TATT alignment amplitudes A₁, A₂, and A_1δ with halo mass, M_h, (in our case, we use M_200mean as the mass of the halo) for our central galaxy sample. Error bars on the bias parameter and amplitude were calculated by taking the 68th percentile values of the IA posteriors. Overplotted in green is the prediction for A₂ and A_1δ from our TATT-M model fit with error bars. Informed by the fits to the halo sample (Fig. A.3), we assume the relations in Eqs. 23 and 24 for the galaxy sample. In orange is the fit to the galaxy sample directly, which we refer to as TATT-M′.