The High Temperature Test Reactor (HTTR) is a graphite-moderated thermal reactor with a nominal power output of 30 MWth. The purpose of this work is to prepare a set of neutron cross sections for use in the transient analyses of three Loss of Forced Cooling (LOFC) events planned for this reactor. The cross sections are generated with the Serpent 2 3-D Monte Carlo code to include the axial heterogeneity and the strong axial coupling of this core. The approach yields good agreement of results when compared to previously used 2-D lattice models, which require a fine spatial discretization and, surprisingly, comes at a higher computational cost. In addition, the temperature distribution must be considered to obtain reasonable cross sections and the axial leakage component dominates the spectral effects in the core. Consequently, a temperature model as a function of the fuel and moderator temperatures is adopted, and the core is depleted to 390 effective full-power days (EFPDs). A full tabulation of cross sections is prepared at this burnup point for a variety of fuel and moderator temperatures. Next, the multiphysics reactor application MAMMOTH is used to evaluate the quality of these cross sections using a diffusion solver. As expected, the homogenization error in the cross sections is significant and the Super Homogenization (SPH) correction from MAMMOTH is necessary to preserve key reaction rates. The SPH-corrected MAMMOTH results are in excellent agreement with the Serpent results and reproduce both the reference power profile and temperature coefficients at each tabulation point. This fact confirms the accuracy of the SPH-corrected cross sections and demonstrates the maturity of using Serpent 2 and MAMMOTH for 3-D cross-section generation, even for reactors as complex as the HTTR. Finally, the authors recommend that the cross sections should be parametrized with the local burnup and that a full core burnup calculation with coupled thermal-fluids would provide a better estimate of the initial condition for any subsequent transient analysis. Unfortunately, this was not within the scope of this work, and direct analysis with these tabulations should be considered as a first-order approximation, at best.