Evaluating Cross-Lingual Unlearning in Multilingual Language Models
We present the first comprehensive evaluation of cross-lingual unlearning in multilingual LLMs. Using translated TOFU benchmarks in seven language/script variants, we test major unlearning algorithms and show that most fail to remove facts outside the training language, even when utility remains high. However, subspace-projection consistently outperforms the other methods, achieving strong cross-lingual forgetting with minimal degradation. Analysis of learned task subspaces reveals a shared interlingua structure: removing this shared subspace harms all languages, while removing language-specific components selectively affects one. These results demonstrate that multilingual forgetting depends on geometry in weight space, motivating subspace-based approaches for future unlearning systems.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Multilingual Amnesia: On the Transferability of Unlearning in Multilingual LLMs
As multilingual large language models become more widely used, ensuring their safety and fairness across diverse linguistic contexts presents unique challenges. While existing research on machine unlearning has primarily…
FAME: Fictional Actors for Multilingual Erasure
LLMs trained on web-scale data raise concerns about privacy and the right to be forgotten. To address these issues, Machine Unlearning provides techniques to remove specific information from trained models without retrai…
Cross-Lingual Unlearning of Selective Knowledge in Multilingual Language Models
Pretrained language models memorize vast amounts of information, including private and copyrighted data, raising significant safety concerns. Retraining these models after excluding sensitive data is prohibitively expens…
Machine UnlearningKnowledge Beyond Language: Bridging the Gap in Multilingual Machine Unlearning Evaluation
While LLMs are increasingly used in commercial services, they pose privacy risks such as leakage of sensitive personally identifiable information (PII). For LLMs trained on multilingual corpora, Multilingual Machine Unle…
Learn and Unlearn in Multilingual LLMs
This paper investigates the propagation of harmful information in multilingual large language models (LLMs) and evaluates the efficacy of various unlearning methods. We demonstrate that fake information, regardless of th…