SLIM: Let LLM Learn More and Forget Less with Soft LoRA and Identity Mixture

TL;DR

SLIM introduces a novel mixture of expert framework combining Soft LoRA and Identity Mixture to enable efficient fine-tuning of large language models, reducing forgetting while maintaining high downstream task performance.

Contribution

The paper proposes SLIM, a mixture of experts approach with dynamic routing and fast adapter merging, to improve PEFT by balancing task adaptation and retention of general capabilities.

Findings

SLIM achieves comparable performance to state-of-the-art PEFT methods.

SLIM significantly mitigates catastrophic forgetting.

SLIM maintains model generality while adapting to downstream tasks.

Abstract

Although many efforts have been made, it is still a challenge to balance the training budget, downstream performance, and the general capabilities of the LLMs in many applications. Training the whole model for downstream tasks is expensive, and could easily result in catastrophic forgetting. By introducing parameter-efficient fine-tuning (PEFT), the training cost could be reduced, but it still suffers from forgetting, and limits the learning on the downstream tasks. To efficiently fine-tune the LLMs with less limitation to their downstream performance while mitigating the forgetting of general capabilities, we propose a novel mixture of expert (MoE) framework based on Soft LoRA and Identity Mixture (SLIM), that allows dynamic routing between LoRA adapters and skipping connection, enables the suppression of forgetting. We adopt weight-yielding with sliding clustering for better out-of-domain distinguish to enhance the routing. We also propose to convert the mixture of low-rank adapters to the model merging formulation and introduce fast dynamic merging of LoRA adapters to keep the general capabilities of the base model. Extensive experiments demonstrate that the proposed SLIM is comparable to the state-of-the-art PEFT approaches on the downstream tasks while achieving the leading performance in mitigating catastrophic forgetting.