mlsanityData to 7 Oct 2026

Multilingual Amnesia: On the Transferability of Unlearning in Multilingual LLMs

arXiv 2601.05641 ↗First posted 9 Jan 2026cs.CL cs.LGcited by 5

On the map: LLM safety and security → Machine unlearning

Abstract (arXiv)

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 focused on monolingual settings, typically English, multilingual environments introduce additional complexities due to cross-lingual knowledge transfer and biases embedded in both pretraining and fine-tuning data. In this work, we study multilingual unlearning using the Aya-Expanse 8B model under two settings: (1) data unlearning and (2) concept unlearning. We extend benchmarks for factual knowledge and stereotypes to ten languages through translation: English, French, Arabic, Japanese, Russian, Farsi, Korean, Hindi, Hebrew, and Indonesian. These languages span five language families and a wide range of resource levels. Our experiments show that unlearning in high-resource languages is generally more stable, with asymmetric transfer effects observed between typologically related languages. Furthermore, our analysis of linguistic distances indicates that syntactic similarity is the strongest predictor of cross-lingual unlearning behavior.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Placed in Machine unlearning, one of 12 directions in LLM safety and security. Builds on 23 of its 39 references that are in the map. First posted on arXiv 9 Jan 2026.

Builds on

23 of 39 references.

Closest ancestors · 7

Same direction, newest first

  1. 2024Cross-Lingual Unlearning of Selective Knowledge in Multilingual Language Models18 Jun
  2. 2024Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning8 Apr
  3. 2024TOFU: A Task of Fictitious Unlearning for LLMs11 Jan
  4. 2023Unlearn What You Want to Forget: Efficient Unlearning for LLMs31 Oct
  5. 2023Fast Model Debias with Machine Unlearning19 Oct
  6. 2023Who's Harry Potter? Approximate Unlearning in LLMs3 Oct
  7. 2022Continual Learning and Private Unlearning24 Mar

Other references · 16

  1. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  2. 2022Locating and Editing Factual Associations in GPTcited by 2,609
  3. 2020mT5: A massively multilingual pre-trained text-to-text transformercited by 1,813
  4. 2019Machine Unlearningcited by 1,020
  5. 2020The State and Fate of Linguistic Diversity and Inclusion in the NLP Worldcited by 789
  6. 2019Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networkscited by 599
  7. 2024Aya Model: An Instruction Finetuned Open-Access Multilingual Language Modelcited by 292
  8. 2023SeeGULL: A Stereotype Benchmark with Broad Geo-Cultural Coverage Leveraging Generative Modelscited by 66
  9. 2023Ethical Reasoning over Moral Alignment: A Case and Framework for In-Context Ethical Policies in LLMscited by 64
  10. 2025Multilingual != Multicultural: Evaluating Gaps Between Multilingual Capabilities and Cultural Alignment in LLMscited by 39
  11. 2025Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Studycited by 38
  12. 2023Investigating Subtler Biases in LLMs: Ageism, Beauty, Institutional, and Nationality Bias in Generative Modelscited by 32
4 more
  1. 2023Language Versatilists vs. Specialists: An Empirical Revisiting on Multilingual Transfer Abilitycited by 27
  2. 2024Translating Across Cultures: LLMs for Intralingual Cultural Adaptationcited by 16
  3. 2024How Well Do LLMs Identify Cultural Unity in Diversity?cited by 8
  4. 2024Evaluating Cultural Awareness of LLMs for Yoruba, Malayalam, and Englishcited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026Multilinguality at the Edge: Developing Language Models for the Global Southcited by 1
  2. 2026Beyond Cross-Lingual Transfer: Benchmarking Propagation Boundaries in Multilingual LLM Unlearningcited by 1
  3. 2026QR-Erase: Efficient Subspace-Based Machine Unlearning with Layer Localizationcited by 0
  4. 2026μ2-Bench: A Multilingual Machine Unlearning Benchmarkcited by 0
  5. 2026Linguistic Loopholes in LLM Unlearning: From a 174-Language Benchmark to Coverage-Aware Unlearningcited by 0