Parallel Universes, Parallel Languages: A Comprehensive Study on LLM-based Multilingual Counterfactual Example Generation

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Wang, Qianli, Nguyen, Van Bach, Liu, Yihong, Splitt, Fedor, Feldhus, Nils, Seifert, Christin, Schütze, Hinrich, Möller, Sebastian, Schmitt, Vera
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917384760066048
author Wang, Qianli
Nguyen, Van Bach
Liu, Yihong
Splitt, Fedor
Feldhus, Nils
Seifert, Christin
Schütze, Hinrich
Möller, Sebastian
Schmitt, Vera
author_facet Wang, Qianli
Nguyen, Van Bach
Liu, Yihong
Splitt, Fedor
Feldhus, Nils
Seifert, Christin
Schütze, Hinrich
Möller, Sebastian
Schmitt, Vera
contents Counterfactuals refer to minimally edited inputs that cause a model's prediction to change, serving as a promising approach to explaining the model's behavior. Large language models (LLMs) excel at generating English counterfactuals and demonstrate multilingual proficiency. However, their effectiveness in generating multilingual counterfactuals remains unclear. To this end, we conduct a comprehensive study on multilingual counterfactuals. We first conduct automatic evaluations on both directly generated counterfactuals in the target languages and those derived via English translation across six languages. Although translation-based counterfactuals offer higher validity than their directly generated counterparts, they demand substantially more modifications and still fall short of matching the quality of the original English counterfactuals. Second, we find the patterns of edits applied to high-resource European-language counterfactuals to be remarkably similar, suggesting that cross-lingual perturbations follow common strategic principles. Third, we identify and categorize four main types of errors that consistently appear in the generated counterfactuals across languages. Finally, we reveal that multilingual counterfactual data augmentation (CDA) yields larger model performance improvements than cross-lingual CDA, especially for lower-resource languages. Yet, the imperfections of the generated counterfactuals limit gains in model performance and robustness.
format Preprint
id arxiv_https___arxiv_org_abs_2601_00263
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Parallel Universes, Parallel Languages: A Comprehensive Study on LLM-based Multilingual Counterfactual Example Generation
Wang, Qianli
Nguyen, Van Bach
Liu, Yihong
Splitt, Fedor
Feldhus, Nils
Seifert, Christin
Schütze, Hinrich
Möller, Sebastian
Schmitt, Vera
Computation and Language
Artificial Intelligence
Counterfactuals refer to minimally edited inputs that cause a model's prediction to change, serving as a promising approach to explaining the model's behavior. Large language models (LLMs) excel at generating English counterfactuals and demonstrate multilingual proficiency. However, their effectiveness in generating multilingual counterfactuals remains unclear. To this end, we conduct a comprehensive study on multilingual counterfactuals. We first conduct automatic evaluations on both directly generated counterfactuals in the target languages and those derived via English translation across six languages. Although translation-based counterfactuals offer higher validity than their directly generated counterparts, they demand substantially more modifications and still fall short of matching the quality of the original English counterfactuals. Second, we find the patterns of edits applied to high-resource European-language counterfactuals to be remarkably similar, suggesting that cross-lingual perturbations follow common strategic principles. Third, we identify and categorize four main types of errors that consistently appear in the generated counterfactuals across languages. Finally, we reveal that multilingual counterfactual data augmentation (CDA) yields larger model performance improvements than cross-lingual CDA, especially for lower-resource languages. Yet, the imperfections of the generated counterfactuals limit gains in model performance and robustness.
title Parallel Universes, Parallel Languages: A Comprehensive Study on LLM-based Multilingual Counterfactual Example Generation
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2601.00263