Re3: Generating Longer Stories With Recursive Reprompting and Revision

By Kevin Yang, Nanyun Peng, Yuandong Tian, Daniel Klein. 2022

274 citations recorded by the source corpus.

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%).

Subfields: Artificial Intelligence, General Social Sciences, Large Language Models, Natural Language Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In General Social Sciences

In Large Language Models

In Natural Language Processing

Citations in the corpus

Self-Refine: Iterative Refinement with Self-Feedback (2023, methodology): “, 2023), Re(3) (Yang et al., 2022), Reflexion (Shinn et al.”
Augmented Language Models: a Survey (2023, methodology): “Re3 (Yang et al., 2022c) exploits this idea to automatically generate stories of over two thousand words.”
When large language models meet personalization: perspectives of challenges and opportunities (2023, unknown): “Re3[218] LLM gpt3-instruct-175B gpt3-instruct-13B Long Stories Generation”
Instruction Tuning for Large Language Models: A Survey (2023, unknown): “LLMs such as GPT-3 [13], PaLM [22], and LLaMA [113] have demonstrated impressive capabilities across a wide range of natural language tasks [1, 19, 38, 40, 41, 59, 60, 75, 76, 91, 104–106, 115, 117, 118, 126, 127, 135, 136, 150].”
Help me write a Poem - Instruction Tuning as a Vehicle for Collaborative Poetry Writing (2022, background): “, 2022) and longer stories (Yang et al., 2022) by generating text that incorporates structural context via prompt chaining.”
Unleashing the potential of prompt engineering in Large Language Models: a comprehensive review (2023, background): “Recursive Reprompting and Revision framework ( Re 3 ) [119] employs zero-shot prompting [49] with GPT-3 to craft a foundational plan encompassing elements such as settings, characters, and outlines.”
Contrastive Distillation Is a Sample-Efficient Self-Supervised Loss Policy for Transfer Learning (2022, unknown): “Similarly, hand-designed actor-critic-like transition functions have been applied to text editing and story-writing to iteratively update contexts based on learned transition functions (Yang et al., 2022; Schick et al., 2022).”
CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquing (2023, background): “, 2022a) ICL External Knowledge NL ✗ (post-hoc) High ✓ Re(3) (Yang et al., 2022) SL & ICL Trained reranker Scalar ✓ (post-hoc) High ✗ LLM-Augmenter (Peng et al.”
A Survey of Large Language Models (2023, unknown): “human writing process to break down language generation into multiple steps such as planning, drafting, rewriting, and editting [342].”
QAmeleon: Multilingual QA with Only 5 Examples (2022, background): “…inter alia ) has further shown that prompting pre-trained large language models (PLMs) can lead to strong performance on various tasks, including question answering (Khashabi et al., 2020; Chowdhery et al., 2022) and open-ended natural language generation (Tang et al., 2022; Yang et al., 2022).”