StyleAdaptedLM：通过高效风格迁移增强指令遵循模型
StyleAdaptedLM: Enhancing Instruction Following Models with Efficient Stylistic Transfer
AI总结：
本文提出 StyleAdaptedLM 框架，通过在非结构化风格语料上训练 LoRA 适配器并与指令模型合并，在不损失指令遵循能力的前提下高效实现品牌或作者风格定制。
摘要：
使大型语言模型（LLM）适应特定风格特征（如品牌语调或作者语气）对于企业沟通至关重要，但要从缺乏指令-响应格式的语料库中实现这一点，同时又不损害指令遵循能力，颇具挑战。我们提出 StyleAdaptedLM，这是一个利用低秩适应（LoRA）将风格特征高效迁移到指令遵循模型的框架。LoRA 适配器首先在基础模型上使用多样化的非结构化风格语料库进行训练，然后与独立的指令遵循模型合并。这使得无需配对数据或牺牲任务性能即可实现稳健的风格定制。跨多个数据集和模型的实验表明，风格一致性得到提升，同时指令遵循能力得以保持，人工评估也证实模型吸收了品牌特有的惯例。StyleAdaptedLM 为 LLM 的风格个性化提供了一条高效路径。
英文摘要：
Adapting LLMs to specific stylistic characteristics, like brand voice or authorial tones, is crucial for enterprise communication but challenging to achieve from corpora which lacks instruction-response formatting without compromising instruction adherence. We introduce StyleAdaptedLM, a framework that efficiently transfers stylistic traits to instruction-following models using Low-Rank Adaptation (LoRA). LoRA adapters are first trained on a base model with diverse unstructured stylistic corpora, then merged with a separate instruction-following model. This enables robust stylistic customization without paired data or sacrificing task performance. Experiments across multiple datasets and models demonstrate improved stylistic consistency while preserving instruction adherence, with human evaluations confirming brand-specific convention uptake. StyleAdaptedLM offers an efficient path for stylistic personalization in LLMs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
