Mix Data or Merge Models? Optimizing for Diverse Multi-Task Learning

TL;DR

This paper compares data mixing and model merging strategies for multilingual multi-task learning, finding that objective-based and language-based merging improve performance and safety more effectively.

Contribution

It demonstrates that model merging, especially objective-based and language-specific merging, outperforms data mixing in enhancing multilingual model safety and performance.

Findings

Objective-based merging improves safety by up to 10%.

Language-based merging increases performance by 4%.

Merging approaches enhance multilingual model safety and effectiveness.

Abstract

Large Language Models (LLMs) have been adopted and deployed worldwide for a broad variety of applications. However, ensuring their safe use remains a significant challenge. Preference training and safety measures often overfit to harms prevalent in Western-centric datasets, and safety protocols frequently fail to extend to multilingual settings. In this work, we explore model merging in a diverse multi-task setting, combining safety and general-purpose tasks within a multilingual context. Each language introduces unique and varied learning challenges across tasks. We find that objective-based merging is more effective than mixing data, with improvements of up to 8% and 10% in general performance and safety respectively. We also find that language-based merging is highly effective -- by merging monolingually fine-tuned models, we achieve a 4% increase in general performance and 7% reduction in harm across all languages on top of the data mixtures method using the same available data. Overall, our comprehensive study of merging approaches provides a useful framework for building strong and safe multilingual models.