Eq. (3.6) and Sec. 3.1: initial Hubble rate, $V(\Phi_i)$, slow-roll parameters, mass and coupling for the benchmark parameters

Work in Planck units $m_{\rm pl}=1$, so $\kappa^2=8\pi$. Take the benchmark parameters of Section 3.1:

the potential $V(\Phi)=v_0+\Lambda^4(1-\Phi^2/\sigma^2)^2$ of (3.1), and the slow-roll value of the initial Hubble rate (3.6),

Then each of the following quantities agrees with the value printed in the paper to within $5\%$ relative error:

1. $H^2(t_i)\simeq9.04\times10^{-19}$;
2. $t_i=H(t_i)^{-1}\simeq5.7\times10^{-35}\,$s, using $m_{\rm pl}=1.86\times10^{43}\,{\rm s}^{-1}$;
3. $V(\Phi_i)\simeq7.2\times10^{-21}$;
4. $\epsilon=\dfrac1{\kappa^2}\left(\dfrac{V'(\Phi_i)}{V(\Phi_i)}\right)^2\simeq0.24$ (slow-roll parameter (1.6));
5. $|\eta|=\left|\dfrac1{\kappa^2}\dfrac{V''(\Phi_i)}{(1+\kappa^2\Phi_i^2/8)V(\Phi_i)}\right|\simeq0.011$ (slow-roll parameter (3.5));
6. $m=\sqrt2\,\mu=\sqrt2\cdot2\Lambda^2/\sigma\simeq7.1\times10^{-11}$;
7. $m\simeq8.6\times10^{8}\,$GeV, using $m_{\rm pl}=1.22\times10^{19}\,$GeV;

and the self-coupling is exactly $\alpha=4\Lambda^4/\sigma^4=2.5\times10^{-21}$.

These are the analytic estimates that fix the benchmark model before the numerical integration.

Formalization Note The paper writes "$\simeq$" with two or three significant digits (and truncates $V(\Phi_i)=7.27\ldots\times10^{-21}$ to $7.2$); a uniform $5\%$ relative tolerance is used to encode "$\simeq$". The e-fold number $N=93$ and the other outputs of the numerical integration are not part of this item.