Evaluating the Ability of Explanations to Disambiguate Models in a Rashomon Set