Low-rank finetuning for LLMs: A fairness perspective

TL;DR

This paper examines the limitations of low-rank fine-tuning methods for large language models, revealing that they may inadequately address dataset shifts and inadvertently preserve biases and toxicity, impacting fairness and safety.

Contribution

It provides empirical evidence that low-rank fine-tuning can fail to mitigate biases and toxicity, highlighting the need for careful evaluation in fairness-sensitive applications.

Findings

Low-rank fine-tuning may not effectively learn dataset shifts.

It can preserve undesirable biases and toxic behaviors.

Implications for fairness and safety in LLM deployment.

Abstract

Low-rank approximation techniques have become the de facto standard for fine-tuning Large Language Models (LLMs) due to their reduced computational and memory requirements. This paper investigates the effectiveness of these methods in capturing the shift of fine-tuning datasets from the initial pre-trained data distribution. Our findings reveal that there are cases in which low-rank fine-tuning falls short in learning such shifts. This, in turn, produces non-negligible side effects, especially when fine-tuning is adopted for toxicity mitigation in pre-trained models, or in scenarios where it is important to provide fair models. Through comprehensive empirical evidence on several models, datasets, and tasks, we show that low-rank fine-tuning inadvertently preserves undesirable biases and toxic behaviors. We also show that this extends to sequential decision-making tasks, emphasizing the need for careful evaluation to promote responsible LLMs development.