Papers by Jordi Cabot
Testing Low-Resource Language Support in LLMs Using Language Proficiency Exams: the Case of Luxembourgish (2026.findings-eacl)
Copied to clipboard
| Challenge: | Large Language Models (LLMs) are used in research and society at large, but are mostly developed with English-speaking users in mind. |
| Approach: | They investigate the viability of language proficiency exams as evaluation tools for Luxembourgish . large models such as Claude and DeepSeek-R1 typically achieve high scores . |
| Outcome: | The proposed models can predict performance in Luxembourgish language tests. |