Advantageous Parameter Expansion Training Makes Better Large Language Models

TL;DR

This paper introduces APEX, a training method that expands advantageous parameters in large language models, leading to better performance with fewer trainable parameters and training data.

Contribution

The paper proposes APEX, a novel parameter expansion training technique that improves large language model training efficiency and performance by focusing on advantageous parameters.

Findings

APEX outperforms full-parameter tuning with only 52% of parameters.

APEX achieves similar perplexity with 33% of training data.

APEX improves downstream task performance.

Abstract

Although scaling up the number of trainable parameters in both pre-training and fine-tuning can effectively improve the performance of large language models, it also leads to increased computational overhead. When delving into the parameter difference, we find that a subset of parameters, termed advantageous parameters, plays a crucial role in determining model performance. Further analysis reveals that stronger models tend to possess more such parameters. In this paper, we propose Advantageous Parameter EXpansion Training (APEX), a method that progressively expands advantageous parameters into the space of disadvantageous ones, thereby increasing their proportion and enhancing training effectiveness. Further theoretical analysis from the perspective of matrix effective rank explains the performance gains of APEX. Extensive experiments on both instruction tuning and continued pre-training demonstrate that, in instruction tuning, APEX outperforms full-parameter tuning while using only 52% of the trainable parameters. In continued pre-training, APEX achieves the same perplexity level as conventional training with just 33% of the training data, and yields significant improvements on downstream tasks.