Corner plot of the posterior for the joint fit of galaxy clustering using non-linear bias and alignment signal using two models: TATT and TATT-M; over the scales 5–100 Mpc/h. TATT-M exploits empirically derived fitting functions between the TATT parameters and halo mass. The galaxy sample consists only of central galaxies, and the sampling was done by fitting b₁, b₂, A₁, A₂, and A_1δ for TATT, and for TATT-M by replacing A₂ and A_1δ with k₁, k₂, and k₃. These three extra parameters are sampled from the posterior of the fits to our halo sample, for which we have better statistics than for the galaxy sample. Due to the large amount of prior information, the overall constraining power is higher with our TATT-M model than with TATT. The posteriors on A₂ and A_1δ for TATT-M are derived using the relations in Eqs. 23 and 24. Note: these were not sampled and are shown here only for comparison.