MetaGPT: Merging Large Language Models Using Model Exclusive Task Arithmetic

TL;DR

MetaGPT introduces a novel model-exclusive task arithmetic method for merging large language models, enhancing multi-task performance while ensuring data privacy and computational efficiency.

Contribution

It formalizes model merging as a multi-task learning problem and leverages LLM properties to create a data-agnostic, cost-effective merging approach.

Findings

Achieves state-of-the-art performance on multiple tasks

Improves task arithmetic effectiveness

Bypasses heavy search processes

Abstract

The advent of large language models (LLMs) like GPT-4 has catalyzed the exploration of multi-task learning (MTL), in which a single model demonstrates proficiency across diverse tasks. Task arithmetic has emerged as a cost-effective approach for MTL. It enables performance enhancement across multiple tasks by adding their corresponding task vectors to a pre-trained model. However, the current lack of a method that can simultaneously achieve optimal performance, computational efficiency, and data privacy limits their application to LLMs. In this paper, we propose \textbf{M}odel \textbf{E}xclusive \textbf{T}ask \textbf{A}rithmetic for merging \textbf{GPT}-scale models, which formalizes the objective of model merging into a multi-task learning framework, aiming to minimize the average loss difference between the merged model and each individual task model. Since data privacy limits the use of multi-task training data, we leverage LLMs' local linearity and task vectors' orthogonality to separate the data term and scaling coefficients term and derive a model-exclusive task arithmetic method. Our proposed MetaGPT is data-agnostic and bypasses the heavy search process, making it cost-effective and easy to implement for LLMs.Extensive experiments demonstrate that MetaGPT leads to improvements in task arithmetic and achieves state-of-the-art performance on multiple tasks.