Computational linguistics

In a bigram language model, which context is used to estimate the probability of the next token?