Interpretable Language Modeling via Induction-head Ngram Models

Re-assign community