Position: Building Guardrails for Large Language Models Requires Systematic Design

Yi Dong, Ronghui Mu, Gaojie Jin, Yi Qi, Jinwei Hu, Xingyu Zhao, Jie Meng, Wenjie Ruan, Xiaowei Huang
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:11375-11394, 2024.

Abstract

As Large Language Models (LLMs) become more integrated into our daily lives, it is crucial to identify and mitigate their risks, especially when the risks can have profound impacts on human users and societies. Guardrails, which filter the inputs or outputs of LLMs, have emerged as a core safeguarding technology. This position paper takes a deep look at current open-source solutions (Llama Guard, Nvidia NeMo, Guardrails AI), and discusses the challenges and the road towards building more complete solutions. Drawing on robust evidence from previous research, we advocate for a systematic approach to construct guardrails for LLMs, based on comprehensive consideration of diverse contexts across various LLMs applications. We propose employing socio-technical methods through collaboration with a multi-disciplinary team to pinpoint precise technical requirements, exploring advanced neural-symbolic implementations to embrace the complexity of the requirements, and developing verification and testing to ensure the utmost quality of the final product.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-dong24c, title = {Position: Building Guardrails for Large Language Models Requires Systematic Design}, author = {Dong, Yi and Mu, Ronghui and Jin, Gaojie and Qi, Yi and Hu, Jinwei and Zhao, Xingyu and Meng, Jie and Ruan, Wenjie and Huang, Xiaowei}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {11375--11394}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/dong24c/dong24c.pdf}, url = {https://proceedings.mlr.press/v235/dong24c.html}, abstract = {As Large Language Models (LLMs) become more integrated into our daily lives, it is crucial to identify and mitigate their risks, especially when the risks can have profound impacts on human users and societies. Guardrails, which filter the inputs or outputs of LLMs, have emerged as a core safeguarding technology. This position paper takes a deep look at current open-source solutions (Llama Guard, Nvidia NeMo, Guardrails AI), and discusses the challenges and the road towards building more complete solutions. Drawing on robust evidence from previous research, we advocate for a systematic approach to construct guardrails for LLMs, based on comprehensive consideration of diverse contexts across various LLMs applications. We propose employing socio-technical methods through collaboration with a multi-disciplinary team to pinpoint precise technical requirements, exploring advanced neural-symbolic implementations to embrace the complexity of the requirements, and developing verification and testing to ensure the utmost quality of the final product.} }
Endnote
%0 Conference Paper %T Position: Building Guardrails for Large Language Models Requires Systematic Design %A Yi Dong %A Ronghui Mu %A Gaojie Jin %A Yi Qi %A Jinwei Hu %A Xingyu Zhao %A Jie Meng %A Wenjie Ruan %A Xiaowei Huang %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-dong24c %I PMLR %P 11375--11394 %U https://proceedings.mlr.press/v235/dong24c.html %V 235 %X As Large Language Models (LLMs) become more integrated into our daily lives, it is crucial to identify and mitigate their risks, especially when the risks can have profound impacts on human users and societies. Guardrails, which filter the inputs or outputs of LLMs, have emerged as a core safeguarding technology. This position paper takes a deep look at current open-source solutions (Llama Guard, Nvidia NeMo, Guardrails AI), and discusses the challenges and the road towards building more complete solutions. Drawing on robust evidence from previous research, we advocate for a systematic approach to construct guardrails for LLMs, based on comprehensive consideration of diverse contexts across various LLMs applications. We propose employing socio-technical methods through collaboration with a multi-disciplinary team to pinpoint precise technical requirements, exploring advanced neural-symbolic implementations to embrace the complexity of the requirements, and developing verification and testing to ensure the utmost quality of the final product.
APA
Dong, Y., Mu, R., Jin, G., Qi, Y., Hu, J., Zhao, X., Meng, J., Ruan, W. & Huang, X.. (2024). Position: Building Guardrails for Large Language Models Requires Systematic Design. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:11375-11394 Available from https://proceedings.mlr.press/v235/dong24c.html.

Related Material