The common alternative to the vector model is the probabilistic model. Probabilistic methods are those which attempt to find a set of documents based on the probability that it satisfies a user's ``Information Need.'' I now look to see how Relevance Feedback is implemented using two probabilistic models. The first is Haines and Croft's implementation of Relevance Feedback in INQUERY using an inference network [16]. The second is a formula derived by Robertson that attempts to select the terms that will have the greatest affect in separating relevant from irrelevant documents [37].