发表机构
Tsinghua University（清华大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出NstAgent，一种无需训练的智能体框架，通过跟踪结构化叙事状态，实现从10K到100K字的长篇故事生成，提升叙事一致性与写作质量。
AI 中文摘要
大语言模型（LLMs）在创意写作方面展现了强大的能力。然而，将其扩展到完整长度的小说仍然具有挑战性，因为维持叙事一致性变得越来越困难。现有的故事生成方法通常专注于约一万字以内的故事，使其扩展到完整长度小说的能力尚未得到充分探索。在这项工作中，我们引入了叙事状态跟踪智能体（Narrative State Tracking Agent，NstAgent），这是一个无需训练的智能体框架，允许LLMs跟踪包括角色、过去事件和未来需求在内的结构化叙事状态。我们扩展了一个现有基准，以比较不同长度下的叙事一致性，并将其与一个写作质量基准一起使用，系统评估从10K到100K字的故事。我们表明，随着故事变长，NstAgent实现了更好的叙事一致性和写作质量，且两者都不会随长度增加而明显下降，这表明它为将故事生成扩展到完整长度小说提供了一种有效的方法。
英文摘要
LLMs have demonstrated strong capabilities in creative writing. However, scaling them to full-length novels remains challenging, as maintaining narrative consistency becomes increasingly difficult. Existing story-generation methods typically focus on stories of up to about ten thousand words, leaving their ability to scale to full-length novels underexplored. In this work, we introduce Narrative State Tracking Agent (NstAgent), a training-free agentic framework that allows LLMs to track a structured narrative state including characters, past events and future requirements. We extend an existing benchmark to compare narrative consistency across lengths, and use it together with a writing-quality benchmark to systematically evaluate stories ranging from 10K to 100K words. We show that NstAgent achieves better narrative consistency and writing quality as stories grow longer, and neither of them degrades noticeably as length increases, suggesting that it provides an effective approach to scaling story generation toward full-length novels.
CommentsUnder review. Code and data are available at https://github.com/zhennan1/NstAgent
