UniversalCEFR: Enabling Open Multilingual Research on Language Proficiency Assessment
Abstract
We introduce UniversalCEFR, a large-scale multilingual and multidimensional dataset of texts annotated with CEFR (Common European Framework of Reference) levels in 13 languages. To enable open research in automated readability and language proficiency assessment, UniversalCEFR comprises 505,807 CEFR-labeled texts curated from educational and learner-oriented resources, standardized into a unified data format to support consistent processing, analysis, and modelling across tasks and languages. To demonstrate its utility, we conduct benchmarking experiments using three modelling paradigms: a) linguistic feature-based classification, b) fine-tuning pre-trained LLMs, and c) descriptor-based prompting of instruction-tuned LLMs. Our results support using linguistic features and fine-tuning pretrained models in multilingual CEFR level assessment. Overall, UniversalCEFR aims to establish best practices in data distribution for language proficiency research by standardising dataset formats, and promoting their accessibility to the global research community.
Cite
@article{arxiv.2506.01419,
title = {UniversalCEFR: Enabling Open Multilingual Research on Language Proficiency Assessment},
author = {Joseph Marvin Imperial and Abdullah Barayan and Regina Stodden and Rodrigo Wilkens and Ricardo Munoz Sanchez and Lingyun Gao and Melissa Torgbi and Dawn Knight and Gail Forey and Reka R. Jablonkai and Ekaterina Kochmar and Robert Reynolds and Eugénio Ribeiro and Horacio Saggion and Elena Volodina and Sowmya Vajjala and Thomas François and Fernando Alva-Manchego and Harish Tayyar Madabushi},
journal= {arXiv preprint arXiv:2506.01419},
year = {2025}
}
Comments
Accepted to EMNLP 2025 (Main Conference)