LoRA-Guard: Parameter-Efficient Guardrail Adaptation for Content Moderation of Large Language Models
Hayder Elesedy, Pedro M Esperanca, Silviu Vlad Oprea, Mete Ozay
Abstract
Guardrails have emerged as an alternative to safety alignment for content moderation of large language models (LLMs). Existing model-based guardrails have not been designed for resource-constrained computational portable devices, such as mobile phones, more and more of which are running LLM-based applications locally. We introduce LoRA-Guard, a parameter-efficient guardrail adaptation method that relies on knowledge sharing between LLMs and guardrail models. LoRA-Guard extracts language features from the LLMs and adapts them for the content moderation task using low-rank adapters, while a dual-path design prevents any performance degradation on the generative task. We show that LoRA-Guard outperforms existing approaches with 100-1000x lower parameter overhead while maintaining accuracy, enabling on-device content moderation.
BibTeX
@inproceedings{elesedy-etal-2024-lora,
title = "{L}o{RA}-Guard: Parameter-Efficient Guardrail Adaptation for Content Moderation of Large Language Models",
author = "Elesedy, Hayder and
Esperanca, Pedro M and
Oprea, Silviu Vlad and
Ozay, Mete",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.656/",
doi = "10.18653/v1/2024.emnlp-main.656",
pages = "11746--11765"
}