Often discrepancies can be found in the corresponding cross sections of different evaluated nuclear data libraries. Traditional integral benchmarks that are used to validate such libraries are sensitive to cross-section values across many different energies. This means an erroneously low cross section at one energy may compensate for an erroneously high cross section at another energy, and the integral benchmark value may still be met. While the evaluated cross section may agree with that single benchmark, it could affect other systems differently. To reduce the potential for this error, an energy differential validation method is proposed herein for continuous energy Monte Carlo neutron transport models in the resolved resonance region and the unresolved resonance region (URR). The proposed method exposes the underlying physics of the URR and validates both the average cross section and resonance self-shielding effect driven by the fluctuations in that cross section. This is done by measuring the neutron transmission of a thick sample that, by its nature, exaggerates the resonance self-shielding effect. This validation method is shown to be very sensitive to the cross-section model used (resolved versus unresolved) and the fluctuation correction employed, allowing it to probe the validity of the previously mentioned cross-section evaluations. Tantalum-181 is used as an example to demonstrate the impact of different resonance evaluations. It was found that the JEFF-3.3 and JENDL-4.0u evaluations made reasonable choices for cross-section models of 181Ta; none of the current evaluations, however, can be used to properly model the validation transmission over all energies. It was also found that updating resonance parameters in the URR provided better agreement with the validation transmission.