arXiv cs.CL· Danrui Li, Yiwei Zhao, Mubbasir Kapadia, Lingxiao Li·· 11 小时前AI 评分29
用轨迹引导故事生成：让 LLM 复现读者阅读动态
Steering Readerly Dynamics: Trajectory-Guided Story Generation with Large Language Models
AI 导读
研究者提出一种从高质量人类故事中学习"读者阅读动态"的框架，用小 LLM 将未来读者反应预测为低维轨迹，再由现成的大型写作 LLM 按这些轨迹生成故事片段。该方法将生成故事对人类写作轨迹的保真度提升 12%-35%，同时基本保持故事连贯性。
正文 · 原文
Abstract:Large language models (LLMs) are increasingly used to generate stories, but these stories struggle to reproduce engaging readerly dynamics, the evolving reader responses that shape narrative experience. To address this challenge, we propose a framework that learns readerly dynamics from high-quality human-written stories and uses them to guide story generation. Instead of directly modeling high-dimensional story content, we train a small LLM to predict future reader responses as low-dimensional trajectories. We then leverage the strong instruction-following capability of a larger, off-the-shelf writer LLM to generate story segments that follow these predicted trajectories within a given story setting. Our framework improves trajectory-based fidelity to human-written stories by 12%-35% while largely preserving story coherence.
| Comments: | 55 pages, 20 figures, 18 tables, including references and the appendix |
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2611.00469 [cs.CL] |
| (or arXiv:2611.00469v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.00469 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Danrui Li [view email]
[v1]
Fri, 30 Oct 2026 23:53:26 UTC (1,058 KB)
来源：arXiv cs.CL · arxiv.org