We examine whether Functional Vectors 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.</p>\n","updatedAt":"2026-09-01T02:54:50.737Z","author":{"_id":"6226f75be1cc4da2221557a1","avatarUrl":"/avatars/6c2b00f27266689b1bb86c4b8d1ac6bb.svg","fullname":"Nguyen Minh Phuong","name":"phuongnm","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.881395697593689},"editors":["phuongnm"],"editorAvatarUrls":["/avatars/6c2b00f27266689b1bb86c4b8d1ac6bb.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2608.29613","authors":[{"_id":"6a963cb6cd6ebc484732ec50","name":"Jieying Xue","hidden":false},{"_id":"6a963cb6cd6ebc484732ec51","name":"Phuong Minh Nguyen","hidden":false},{"_id":"6a963cb6cd6ebc484732ec52","name":"Minh Le Nguyen","hidden":false},{"_id":"6a963cb6cd6ebc484732ec53","name":"Shogo Okada","hidden":false}],"publishedAt":"2026-08-30T00:00:00.000Z","submittedOnDailyAt":"2026-09-01T00:00:00.000Z","title":"Cross-lingual Functional Vectors for Emotion Detection in Large Language Models","submittedOnDailyBy":{"_id":"6226f75be1cc4da2221557a1","avatarUrl":"/avatars/6c2b00f27266689b1bb86c4b8d1ac6bb.svg","isPro":false,"fullname":"Nguyen Minh Phuong","user":"phuongnm","type":"user","name":"phuongnm"},"summary":"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.","upvotes":0,"discussionId":"6a963cb6cd6ebc484732ec54","githubRepo":"https://github.com/yingjie7/cross_lingual_fvs","githubRepoAddedBy":"user","ai_summary":"Function vectors extracted from one language improve multilingual emotion recognition by capturing language-independent task signals and reducing inference overhead.","ai_keywords":["function vectors","large language models","cross-lingual transfer","multilingual emotion recognition","attention heads","in-context learning","zero-shot"],"ai_summary_model":"thinkingmachines/Inkling-Small","githubStars":0},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[],"acceptLanguages":["en"],"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2608/2608.29613.md","query":{}}">
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 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.
Community
We examine whether Functional Vectors 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.
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2608.29613 in a model README.md to link it from this page.
Cite arxiv.org/abs/2608.29613 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2608.29613 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.