Experimental data of the reverse annealing of the effective concentration of ionized space charges (N-eff, also called effective doping or impurity concentration) of neutron irradiated high resistivity silicon detectors has been compared with various models. The models include the compensation model (first order), the cluster model of the first order, the neutral to acceptor model (first order), and the cluster model of the second order. Detectors irradiated to various neutron fluences have been annealed at 80 degrees C for up to 17 hours to reach the saturation of the first apparent stage of the N-eff reverse anneal, which is equivalent of about one year of room temperature (RT) anneal. The annealing time constant, defined as the time at half saturation tau 1/2 has been found to be virtually a constant (similar to 140 minutes+/-14%) for detectors irradiated to fluences ranging from 8.2x10(12) n/cm(2) to 3.2x10(13) n/cm(2), which is the distinguishing characteristic of the first order process. The least square fit of the data to the first order models has shown a time constant of 221.7 minutes with a 14% error and that to the second order model has shown a k constant of 7.3x10(-5) s(-1) with a 37% error. The best fit, however, is a first order dt with two time constants: a short one (similar to 44 minutes+/-2.5%) with a small amplitude and a longer one (similar to 290 minutes+/-l2%) with an almost five times larger amplitude, suggesting that even for the apparent first stage of the N-eff reverse anneal, there may be two stages. There is also evidence that even after the apparent first stage anneal, there is at least another stage which is showing up in higher temperature annealing (150 degrees C).