少样本作者风格迁移：预测 steering 向量与 LoRA 权重的方法研究
Predicting Steering Vectors and Adapter Weights for Few-Shot Author-Style Transfer
教你怎么用几篇摘要就模仿某个作者的写作风格，试了 steering、LoRA 和 hypernetwork 三条路，hypernetwork 效果最均衡，做写作风格化可以参考。
论文研究从每位作者的少量摘要样例中学习写作风格，对比了三种方法：对比激活 steering、预测 steering 向量的网络，以及预测 LoRA 适配器的 hypernetwork。实验显示微调能获取大部分风格信号但损失流畅度，hypernetwork 在已见和未见作者上都取得风格模仿与输出质量的最佳平衡。steering 在作者层面操作，将作者摘要与同内容风格中性的生成结果对比，固定了主题变量，效果优于基于预设风格清单的方案。分析还发现人工提取与预测的 steering 向量接近正交但得分相当，说明风格条件化至少存在两个不相关的方向。
Predicting Steering Vectors and Adapter Weights for Few-Shot Author-Style Transfer
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.
