Do Multilingual Language Models Think Better in English?
This addresses the issue of suboptimal performance in non-English languages for users of multilingual models, though it is incremental as it builds on the translate-test technique.
The paper tackles the problem that multilingual language models underperform in non-English languages, and introduces self-translate, a method that uses the models' own few-shot translation capabilities to translate inputs into English before inference, consistently outperforming direct inference across 5 tasks.
Translate-test is a popular technique to improve the performance of multilingual language models. This approach works by translating the input into English using an external machine translation system, and running inference over the translated input. However, these improvements can be attributed to the use of a separate translation system, which is typically trained on large amounts of parallel data not seen by the language model. In this work, we introduce a new approach called self-translate, which overcomes the need of an external translation system by leveraging the few-shot translation capabilities of multilingual language models. Experiments over 5 tasks show that self-translate consistently outperforms direct inference, demonstrating that language models are unable to leverage their full multilingual potential when prompted in non-English languages. Our code is available at https://github.com/juletx/self-translate.