(Preprint) Improving Bilingual Capabilities of Language Models to Support Diverse Linguistic Practices in Education
To evaluate the reliability of multilingual Large Language Models (MLLM) tools in bilingual classrooms, we use the mixture of English and Spanish, Spanglish, as a representation of bilingual students. This work will be submitted in Spring 2025.
Citation: Syamkumar, Anand, Nora Tseng, Kaycie Barron, Shanglin Yang, Shamya Karumbaiah, Rheeya Uppaal, and Junjie Hu. “Improving Bilingual Capabilities of Language Models to Support Diverse Linguistic Practices in Education.” arXiv preprint arXiv:2411.04308 (2024).