arXiv：cs.CL· Zahra Solati Dehkordi, Vasileios Lampos·· 6 小时前AI 评分33
DivLM：提升 LLM 短篇故事生成多样性
Improving Diversity in LLM Short Story Generation
AI 导读
研究者提出 DivLM，一个两阶段 LLM 后训练框架，用于提升创意短篇故事生成的多样性。该方法先在创意写作语料上继续预训练并用权重残差恢复指令遵循能力，再通过自定义复合奖励函数进行强化学习，联合优化体裁、语气、风格和命名实体等维度的多样性。在两个 LLM 系列上的实验显示，DivLM 的多样性指标平均提升超过 9%，同时保持指令遵循、回复质量和与人类输出的相似度。
正文
Abstract:Large language models (LLMs) can generate accurate responses, but these are void of diversity. We attempt to address this for the task of creative short story generation. Drawing on established writing conventions and known LLM limitations, we target variation in genre, tone, style, and named entities. To promote diversity across these dimensions, we introduce DivLM, an LLM post-training framework consisting of two phases. First, we perform continued pre-training on a creative writing corpus and restore instruction-following capabilities using weight residuals. We then apply reinforcement learning with a custom, composite reward function that jointly maximizes diversity across the targeted narrative dimensions while maintaining response quality. Our empirical results on two LLM families show that DivLM increases diversity metrics by more than 9% on average compared to alternative approaches, while preserving instruction following, overall response quality, and similarity to human outputs.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.06729 [cs.CL] |
| (or arXiv:2610.06729v2 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.06729 arXiv-issued DOI via DataCite |
Submission history
From: Zahra Solati Dehkordi [view email]
[v1]
Mon, 5 Oct 2026 17:16:29 UTC (686 KB)
[v2]
Tue, 6 Oct 2026 17:15:34 UTC (686 KB)
来源：arXiv：cs.CL · arxiv.org