Information Parity: Measuring and Predicting the Multilingual Capabilities of Language Models

Alexander Tsvetkov, Alon Kipnis


Abstract
Large Language Models (LLMs) are increasingly deployed in user-facing applications worldwide, necessitating handling multiple languages across various tasks. We propose a metric called Information Parity (IP) that can predict an LLM’s capabilities across multiple languages in a task-agnostic manner. IP is well-motivated from an information theoretic perspective: it is associated with the LLM’s efficiency of compressing the text in a given language compared to a reference language. We evaluate IP and other popular metrics such as Tokenization Parity (TP) and Tokenizer Fertility (TF) on several variants of open-sourced LLMs (Llama2, Gemma, Mistral). Among all metrics known to us, IP is better correlated with existing task-specific benchmark scores from the literature and thus better predicts such scores in a certain language. These findings show that IP may be useful for ranking multilingual LLMs’ capabilities regardless of the downstream task.
Anthology ID:
2024.findings-emnlp.468
Volume:
Findings of the Association for Computational Linguistics: EMNLP 2024
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
7971–7989
Language:
URL:
https://aclanthology.org/2024.findings-emnlp.468
DOI:
Bibkey:
Cite (ACL):
Alexander Tsvetkov and Alon Kipnis. 2024. Information Parity: Measuring and Predicting the Multilingual Capabilities of Language Models. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 7971–7989, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
Information Parity: Measuring and Predicting the Multilingual Capabilities of Language Models (Tsvetkov & Kipnis, Findings 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.findings-emnlp.468.pdf