arXiv：cs.AI· Leonard Popp, Danni Liu, Supriti Sinhamahapatra, Jan Niehues·· 4 小时前AI 评分30
预测引导向量与适配器权重实现少样本作者风格迁移
Predicting Steering Vectors and Adapter Weights for Few-Shot Author-Style Transfer
AI 导读
研究提出三种从少量摘要样本实现作者风格条件生成的方法：对比激活引导、预测引导向量的网络，以及预测 LoRA 适配器的超网络。结果显示风格模仿与输出质量存在权衡，微调能获取最多风格信号但损失流畅度，超网络在已见和未见作者上均取得最佳平衡。引导在作者层面操作，将同一内容与风格中性生成对比，无需预定义风格清单，且优于基于清单的引导。
正文
Abstract:Adapting large language models to an individual author's style from a few examples is challenging, and scientific writing sharpens the difficulty: formal conventions leave little surface variation, and authors write about their own topics, so extracted ``style'' easily entangles with content. We study style-conditioned abstract generation from a few example abstracts per author and propose three methods: (1) contrastive activation steering, (2) a network that predicts steering vectors, and (3) a hypernetwork that predicts LoRA adapters. We find a consistent trade-off between style imitation and output quality: fine-tuning buys most of the available style signal but forfeits fluency, while the hypernetwork achieves the best trade-off on both seen and unseen authors. Our steering operates at author level, contrasting an author's abstracts against style-neutral generations for the same content. This holds topic fixed, removes the need for a predefined style inventory, and outperforms inventory-based steering. % [EDIT 1a] softened "no single optimal axis" claim Moreover, our analyses demonstrate that manually extracted and predicted steering vectors are near-orthogonal yet score comparably, indicating that style conditioning here can admit at least two unrelated directions rather than requiring one particular axis.
| Comments: | W-NUT Workshop @ EMNLP 2026 |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.03163 [cs.CL] |
| (or arXiv:2610.03163v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.03163 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Danni Liu [view email]
[v1]
Fri, 2 Oct 2026 11:43:14 UTC (377 KB)
来源：arXiv：cs.AI · arxiv.org