LoRE-Merging: Exploring Low-Rank Estimation For Large Language Model Merging

TL;DR

LoRE-Merging introduces a low-rank estimation framework for large language model merging that enhances model performance without additional training, effectively reducing interference and preserving task-specific knowledge.

Contribution

The paper presents a novel low-rank estimation approach for model merging that does not require access to the base model, improving efficiency and effectiveness.

Findings

Effective in mitigating interference between models

Preserves task-specific information better than existing methods

Advances state-of-the-art in model merging techniques

Abstract

While most current approaches rely on further training techniques, such as fine-tuning or reinforcement learning, to enhance model capacities, model merging stands out for its ability of improving models without requiring any additional training. In this paper, we propose a unified framework for model merging based on low-rank estimation of task vectors without the need for access to the base model, named \textsc{LoRE-Merging}. Our approach is motivated by the observation that task vectors from fine-tuned models frequently exhibit a limited number of dominant singular values, making low-rank estimations less prone to interference. We implement the method by formulating the merging problem as an optimization problem. Extensive empirical experiments demonstrate the effectiveness of our framework in mitigating interference and preserving task-specific information, thereby advancing the state-of-the-art performance in model merging techniques.