Poro 34B and the Blessing of Multilinguality

TL;DR

This paper introduces Poro 34B, a multilingual language model trained on Finnish, English, and programming languages, demonstrating that multilingual training can enhance performance for low-resource languages and translation tasks.

Contribution

The study shows that multilingual training with limited data can outperform monolingual models for specific languages, and introduces a new 34B parameter model with open access.

Findings

Outperforms existing models for Finnish and translation tasks

Achieves competitive performance in English and programming languages

Demonstrates benefits of multilingual training with limited data

Abstract

The pretraining of state-of-the-art large language models now requires trillions of words of text, which is orders of magnitude more than available for the vast majority of languages. While including text in more than one language is an obvious way to acquire more pretraining data, multilinguality is often seen as a curse, and most model training efforts continue to focus near-exclusively on individual large languages. We believe that multilinguality can be a blessing: when the lack of training data is a constraint for effectively training larger models for a target language, augmenting the dataset with other languages can offer a way to improve over the capabilities of monolingual models for that language. In this study, we introduce Poro 34B, a 34 billion parameter model trained for 1 trillion tokens of Finnish, English, and programming languages, and demonstrate that a multilingual training approach can produce a model that substantially advances over the capabilities of existing models for Finnish and excels in translation, while also achieving competitive performance in its class for English and programming languages. We release the model parameters, scripts, and data under open licenses at https://huggingface.co/LumiOpen/Poro-34B.