Char-mander Use mBackdoor! A Study of Cross-lingual Backdoor Attacks in Multilingual LLMs
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915532247138304 |
|---|---|
| author | Beniwal, Himanshu Panda, Sailesh Srivibhav, Birudugadda Singh, Mayank |
| author_facet | Beniwal, Himanshu Panda, Sailesh Srivibhav, Birudugadda Singh, Mayank |
| contents | We explore \textbf{C}ross-lingual \textbf{B}ackdoor \textbf{AT}tacks (X-BAT) in multilingual Large Language Models (mLLMs), revealing how backdoors inserted in one language can automatically transfer to others through shared embedding spaces. Using toxicity classification as a case study, we demonstrate that attackers can compromise multilingual systems by poisoning data in a single language, with rare and high-occurring tokens serving as specific, effective triggers. Our findings expose a critical vulnerability that influences the model's architecture, resulting in a concealed backdoor effect during the information flow. Our code and data are publicly available https://github.com/himanshubeniwal/X-BAT. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2502_16901 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Char-mander Use mBackdoor! A Study of Cross-lingual Backdoor Attacks in Multilingual LLMs Beniwal, Himanshu Panda, Sailesh Srivibhav, Birudugadda Singh, Mayank Computation and Language Artificial Intelligence We explore \textbf{C}ross-lingual \textbf{B}ackdoor \textbf{AT}tacks (X-BAT) in multilingual Large Language Models (mLLMs), revealing how backdoors inserted in one language can automatically transfer to others through shared embedding spaces. Using toxicity classification as a case study, we demonstrate that attackers can compromise multilingual systems by poisoning data in a single language, with rare and high-occurring tokens serving as specific, effective triggers. Our findings expose a critical vulnerability that influences the model's architecture, resulting in a concealed backdoor effect during the information flow. Our code and data are publicly available https://github.com/himanshubeniwal/X-BAT. |
| title | Char-mander Use mBackdoor! A Study of Cross-lingual Backdoor Attacks in Multilingual LLMs |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2502.16901 |