The most accurate determination of neutron yields from fusion reactors maybe obtained from neutron activation measurements of elemental foils. On the Tokamak Fusion Test Reactor (TFTR), a re-entrant irradiation end has been installed to provide a low-scattering environment close to the plasma for neutron activation measurements. The ratio of energy-dependent fluence to total fusion yield is calculated using a fully three-dimensional Monte Carlo calculation with the Monte Carlo code for neutron and photon transport (MCNP). Corrections to the “virgin” fluence from attenuation and scattering are only 10 to 20% for deuterium-tritium (D-T) reactions and 30 to 40% for deuterium-deuterium reactions. A total 1-sigma accuracy of ±8% is achieved for D-T neutron yields over a wide dynamic range. This paper documents the response coefficients (hits per source neutron, where hits are activated nuclei per target nuclei) for use by the neutron activation system on TFTR; describes the possible systematic corrections needed (such as major radial variations or the impact of ion temperature on reactions with high-energy thresholds); and estimates uncertainties in the response coefficients. Results from in situ use of a D-T neutron generator are also analyzed using the MCNP modeling as an approximate benchmarking experiment; only 20% accuracy in the comparison is possible because of poor counting statistics in the calibration experiment.