Investigating Multilingual Instruction-Tuning: Do Polyglot Models Demand for Multilingual Instructions?

Alexander Arno Weber, Klaudia Thellmann, Jan Ebert, Nicolas Flores-Herr, Jens Lehmann, Michael Fromm, Mehdi Ali


Abstract
The adaption of multilingual pre-trained LLMs into eloquent and helpful assistants is essential to facilitate their use across different language regions. In that spirit, we are the first to conduct an extensive study of the performance of multilingual models instruction-tuned on different language compositions on parallel instruction-tuning benchmarks across a selection of the most spoken Indo-European languages. We systematically examine the effects of language and instruction dataset size on a mid-sized and a large, multilingual LLMs by instruction-tuning them on parallel instruction-tuning datasets. Our results demonstrate that instruction-tuning on parallel instead of monolingual corpora benefits cross-lingual instruction following capabilities by up to 9.9%. Furthermore, we show that the Superficial Alignment Hypothesis does not hold in general, as the investigated multilingual 7B parameter model presents a counter-example requiring large-scale instruction-tuning datasets. Finally, we conduct a human annotation study to understand the alignment between human-based and GPT-4-based evaluation within multilingual chat scenarios.
Anthology ID:
2024.emnlp-main.1159
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
20829–20855
Language:
URL:
https://aclanthology.org/2024.emnlp-main.1159
DOI:
10.18653/v1/2024.emnlp-main.1159
Bibkey:
Cite (ACL):
Alexander Arno Weber, Klaudia Thellmann, Jan Ebert, Nicolas Flores-Herr, Jens Lehmann, Michael Fromm, and Mehdi Ali. 2024. Investigating Multilingual Instruction-Tuning: Do Polyglot Models Demand for Multilingual Instructions?. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 20829–20855, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Investigating Multilingual Instruction-Tuning: Do Polyglot Models Demand for Multilingual Instructions? (Weber et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-main.1159.pdf
Software:
 2024.emnlp-main.1159.software.zip
Data:
 2024.emnlp-main.1159.data.zip