Language Models Have Interpretable Narrative Mechanisms

We investigate whether language models contain internal components that contribute to representing narrativity, the degree to which a text exhibits narrative properties.