BayLing 2: A Multilingual Large Language Model with Efficient Language Alignment

TL;DR

BayLing 2 is a multilingual large language model that efficiently transfers knowledge from high-resource to low-resource languages using language alignment and instruction tuning, significantly improving performance across over 100 languages.

Contribution

Introduces BayLing 2, a model that leverages language alignment and a large instruction dataset to enhance multilingual capabilities without extensive instruction data for each language.

Findings

Outperforms similar-scale open-source models in multilingual translation.

Achieves significant improvements in low-resource language understanding.

Maintains high performance in high-resource languages while improving low-resource language capabilities.

Abstract

Large language models (LLMs), with their powerful generative capabilities and vast knowledge, empower various tasks in everyday life. However, these abilities are primarily concentrated in high-resource languages, leaving low-resource languages with weaker generative capabilities and relatively limited knowledge. Enhancing the multilingual capabilities of LLMs is therefore crucial for serving over 100 linguistic communities worldwide. An intuitive approach to enhance the multilingual capabilities would be to construct instruction data for various languages, but constructing instruction data for over 100 languages is prohibitively costly. In this paper, we introduce BayLing 2, which efficiently transfers generative capabilities and knowledge from high-resource languages to low-resource languages through language alignment. To achieve this, we constructed a dataset of 3.2 million instructions, comprising high-resource language instructions (Chinese and English) and cross-lingual instructions for 100+ languages and performed instruction tuning based on the dataset to facilitate the capability transfer between languages. Using Llama as the foundation model, we developed BayLing-2-7B, BayLing-2-13B, and BayLing-2-8B, and conducted a comprehensive evaluation of BayLing. For multilingual translation across 100+ languages, BayLing shows superior performance compared to open-source models of similar scale. For multilingual knowledge and understanding benchmarks, BayLing achieves significant improvements across over 20 low-resource languages, demonstrating its capability of effective knowledge transfer from high-resource to low-resource languages. Furthermore, results on English benchmarks indicate that BayLing maintains high performance in highresource languages while enhancing the performance in low-resource languages. Demo, homepage, code and models of BayLing are available.