Notes. The first column shows the ID of each bin. The second column presents the mass range for the halos as taken from the PNG-UNIT (Sect. 4.1). In the third column, we show the measurements we get for the product b_ϕf_NL using the standard errors (σ_S(P(k))) for the Gaussian UNIT simulation. In the fourth column, we present the results of b_ϕf_NL after applying the matching between the PNG-UNIT and the UNIT simulations (Sect. 5.2). The fifth column provides the f_NL values obtained from the PNG-UNIT simulation using standard errors and assuming the universality relation, p = 1, in Eq. (13). The sixth column shows the Pearson correlation coefficient, ρ, assumed for each halo mass bin. These values have been obtained after a linear regression of the quantities measured for bins 3–10, as described in Sect. 4.3. For bins 0–2, we use instead the mean of the measurements of ρ, which is $\overline{\rho }=0.703$, as discussed in the same section. In the seventh column, we present the results when applying the reduced errors derived from the FASTPM mocks and matching them with the Gaussian simulation (Sect. 5.3). The last column displays the values of p obtained from Eq. (13), which are used to correct the value of f_NL in the fifth column to the input value (Sect. 5.4). We find that the error estimates are not robust for the bins in grey and may lead to biases in the conclusions about b_ϕ. Therefore, as discussed in Sect. 5.2, they are not considered for subsequent analyses.