AI 中文总结
针对现有长故事生成方法易累积各类错误的问题，提出无训练框架ConWriter，通过维护故事状态、验证叙事过渡并结合风险信号实现一致性控制，在ConStory-Bench上完成多模型多任务评估。
AI 中文摘要
长文本故事生成要求模型在扩展上下文中保持叙事一致性，但现有基于提示的方法往往会随着故事的发展累积时间、事实、角色、常识和风格错误。我们提出ConWriter，这是一种用于一致性感知长故事生成的无训练框架。ConWriter在场景级别逐步生成故事，受静态故事需求、动态叙事记忆、符号状态推理和不确定性感知风险信号的引导。ConWriter不将长故事生成视为单一的自由格式解码过程，而是维护不断演变的故事状态，检查新场景是否满足所需的叙事过渡，并使用不确定性感知风险信号来优先进行验证和局部修复。这使得在生成过程中就能进行一致性控制，避免局部错误传播到后续场景。我们在ConStory-Bench上对ConWriter进行评估，该基准涵盖四个长故事任务：续写、生成、扩展和补全。由于长文本生成和评估的成本较高，我们采用每个任务的前5个案例，并在Qwen3.5-Plus、DeepSeek-V4-Flash和GPT-5系列上测试3k、6k和12k的目标长度。实验遵循官方ConStory-Bench评估协议。
英文摘要
Long-form story generation requires models to preserve narrative consistency across extended contexts, yet existing prompting-based methods often accumulate temporal, factual, character, commonsense, and stylistic errors as the story grows. We propose ConWriter, a training-free framework for consistency-aware long-form story generation. ConWriter writes stories incrementally at the scene level, guided by static story requirements, dynamic narrative memory, symbolic state reasoning, and uncertainty-aware risk signals. Rather than treating long-story generation as a single free-form decoding process, ConWriter maintains evolving story states, checks whether new scenes satisfy required narrative transitions, and uses uncertainty-aware risk signals to prioritize validation and localized repair. This enables consistency control during generation, before local errors propagate into later scenes. We evaluate ConWriter on ConStory-Bench across four long-story tasks, three target lengths, and multiple base LLMs following the official evaluation protocol. Across models and story lengths, ConWriter consistently matches or improves upon direct generation and outperforms the recent training-free baseline DOME in narrative consistency. These results demonstrate the effectiveness of lightweight neuro-symbolic consistency control for training-free long-form story generation. Code is available on \href{https://github.com/jindongli-Ai/ConWriter}{GitHub}.
CommentsAccepted to Findings of EMNLP 2026
