跨语言功能向量用于大语言模型情感检测
原标题：Paper page - Cross-lingual Functional Vectors for Emotion Detection in Large Language Models
AI 摘要
该研究探讨了功能向量（FVs）在大型语言模型中的跨语言迁移能力，以多语言多标签情感识别为基准。实验表明，从源语言提取的FVs能在零样本设置下有效提升目标语言的任务表现，且不依赖推理时的演示，表明FVs捕获了语言无关的任务信号。研究还发现每个LLM存在稳定的最优注意力头范围，FVs可部分替代少样本上下文学习并降低计算开销。代码已开源。
正文节选
Cross-lingual Functional Vectors for Emotion Detection in Large Language Models Abstract Function vectors extracted from one language improve multilingual emotion recognition by capturing language-independent task signals and reducing inference overhead. 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 studi