Efficient, low-cost, and stable solid-state materials for portable thermal neutron detection are highly expected in order to substitute the currently used 3He and BF3 tank detectors. Li-glasses and halide compounds have emerged as candidates, however, they present critical drawbacks for practical implementations. Ce:Li6Y(BO3)3 is a promising oxide candidate that however has been disregarded due to its negligibly low light yield, caused by a poor crystalline and optical quality. In this study, we demonstrate that a post-growth thermal treatment is the key parameter to drastically reduce the concentration of intrinsic defects and scattering centers that lead to severe non-radiative recombination. Even though this annealing also involves the oxidation of activator Ce3+ ions to Ce4+, a drastic enhancement of the light yield by ~600% is achieved independently of the Ce3+ concentration within the considered range. The light yield of 4650 ph/n is already close to that of commercial GS20 with 6000 ph/n.