Extending LLMs to New Languages: A Case Study of Llama and Persian Adaptation
This work addresses the challenge of extending LLMs to low-resource languages for users in those linguistic communities, but it is incremental as it builds on existing adaptation methods with limited novel insights.
The study tackled the problem of adapting English-centric LLMs to low-resource languages like Persian using parameter-efficient fine-tuning, finding that bilingual data alignment improved Persian classification accuracy without harming English performance, with cross-lingual alignment offering minimal benefits for the low-resource language.
Large language models (LLMs) have made great progress in classification and text generation tasks. However, they are mainly trained on English data and often struggle with low-resource languages. In this study, we explore adding a new language, i.e., Persian, to Llama (a model with a limited understanding of Persian) using parameter-efficient fine-tuning. We employ a multi-stage approach involving pretraining on monolingual Persian data, aligning representations through bilingual pretraining and instruction datasets, and instruction-tuning with task-specific datasets. We evaluate the model's performance at each stage on generation and classification tasks. Our findings suggest that incorporating the Persian language, through bilingual data alignment, can enhance classification accuracy for Persian tasks, with no adverse impact and sometimes even improvements on English tasks. Additionally, the results highlight the model's initial strength as a critical factor when working with limited training data, with cross-lingual alignment offering minimal benefits for the low-resource language. Knowledge transfer from English to Persian has a marginal effect, primarily benefiting simple classification tasks.