Re3: Generating Longer Stories With Recursive Reprompting and Revision

Authors:
  • KYKevin Yang
    Kevin Yang
  • YTYuandong Tian
    Yuandong Tian
  • NPNanyun Peng
    Nanyun Peng
  • DKDan Klein
    Dan Klein

Abstract
We consider the problem of automatically generating longer stories of over two thousand words. Compared to prior work on shorter stories, long-range plot coherence and relevance are more central challenges here. We propose the Recursive Reprompting and Revision framework (Re3) to address these challenges by (a) prompting a general-purpose language model to construct a structured overarching plan, and (b) generating story passages by repeatedly injecting contextual information from both the plan and current story state into a language model prompt. We then revise by (c) reranking different continuations for plot coherence and premise relevance, and finally (d) editing the best continuation for factual consistency. Compared to similar-length stories generated directly from the same base model, human evaluators judged substantially more of Re3's stories as having a coherent overarching plot (by 14% absolute increase), and relevant to the given initial premise (by 20%).