发表机构
Japan Advanced Institute of Science and Technology（日本先进科学技术学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究以多语言多标签情感识别为基准，探究跨语言功能向量（FVs）在大语言模型中的应用，发现FVs可跨语言提升情感检测性能，是轻量可迁移的多语言任务适配机制。
AI 中文摘要
功能向量（Function Vectors，简称FVs）近来成为一种有前景的机制，可通过注入从上下文示例中推导的特定任务潜在方向表示来引导大语言模型（Large Language Models，简称LLMs）的行为。尽管已有研究表明，FVs能在结构化上下文学习场景中恢复任务行为，但它们在语义复杂任务上的有效性及跨语言泛化能力仍未得到充分探索。本研究以多语言多标签情感识别作为具有挑战性的语义分类基准，探究FVs的跨语言可迁移性。具体而言，我们检验在标准干净设置和扰动零样本设置下，推理过程不提供示例时，源语言中提取的FVs是否能引导另一语言的任务行为。在多种跨语言场景中，应用FVs可显著提升性能，表明FVs捕捉到与语言无关、与任务相关的信号，而非纯粹的语言特定词汇模式，凸显其作为轻量且可迁移的多语言任务适配机制的潜力。我们还观察到，每个大语言模型在构建有效FVs时，注意力头存在相对稳定的最优范围，且该模式在不同语言间保持一致。此外，FVs可部分复制标准少样本上下文学习的任务引导效果，同时避免处理多个示例的计算开销，使其适用于大规模实际应用。我们的代码可在https URL获取。
英文摘要
Function vectors (FVs) have recently emerged as a promising mechanism for steering the behavior of large language models (LLMs) by injecting task-specific latent direction representations derived from in-context demonstrations. While prior studies have shown that FVs can recover task behavior in structured in-context learning settings, their effectiveness on semantically complex tasks and their ability to generalize across languages remain underexplored. We investigate the cross-lingual transferability of FVs using multilingual multi-label emotion recognition as a challenging semantic classification benchmark. Specifically, we examine whether FVs extracted from a source language can steer task behavior in another language under both standard clean and perturbed zero-shot settings without providing demonstrations during inference. Across diverse cross-lingual settings, applying FVs substantially improves performance, suggesting that FVs capture language-agnostic, task-relevant signals rather than purely language-specific lexical patterns, and highlighting their potential as a lightweight and transferable mechanism for multilingual task adaptation. We observe that each LLM exhibits a relatively stable optimal range of attention heads for constructing effective FVs, and the pattern remains consistent across languages. In addition, FVs can partially replicate the task-steering effects of standard few-shot in-context learning while avoiding the computational overhead of processing multiple demonstrations, making them effective for large-scale practical applications. Our code is available at https://github.com/yingjie7/cross_lingual_fvs.
CommentsFindings of the Association for Computational Linguistics: EMNLP 2026
