paper-with-me

Papers

Building Guardrails for Large Language Models

2024-02-02 · Yi Dong, Ronghui Mu, Gaojie Jin, Yi Qi, Jinwei Hu, Xingyu Zhao, Jie Meng, Wenjie Ruan, Xiaowei Huang

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.

📄 PDF Abstract BibTeX arXiv:2402.01822

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Triaging Threats to Specialized Guardrails

2026-05-29 · Wenjie Jacky Mo, Xiaofei Wen, Rui Cai, Boyu Zhu 외 arxiv

Building robust safety guardrails is essential for deploying Large Language Models across diverse real-world applications. However, this goal remains challenging because safety risks span heterogeneous threat domains, wh…

Why LLM Safety Guardrails Collapse After Fine-tuning: A Similarity Analysis Between Alignment and Fine-tuning Datasets

2025-06-05 · Lei Hsiung, Tianyu Pang, Yung-Chen Tang, Linyue Song 외

Recent advancements in large language models (LLMs) have underscored their vulnerability to safety alignment jailbreaks, particularly when subjected to downstream fine-tuning. However, existing mitigation strategies prim…

Safety Alignment

OmniGuard: Unified Omni-Modal Guardrails with Deliberate Reasoning

2025-12-02 · Boyu Zhu, Xiaofei Wen, Wenjie Jacky Mo, Tinghui Zhu 외 arxiv

Omni-modal Large Language Models (OLLMs) that process text, images, videos, and audio introduce new challenges for safety and value guardrails in human-AI interaction. Prior guardrail research largely targets unimodal se…

Binary Classification

Breaking Guardrails, Facing Walls: Insights on Adversarial AI for Defenders & Researchers

2025-10-14 · Giacomo Bertollo, Naz Bodemir, Jonah Burgess arxiv

Analyzing 500 CTF participants, this paper shows that while participants readily bypassed simple AI guardrails using common techniques, layered multi-step defenses still posed significant challenges, offering concrete in…

Benchmarking LLM Guardrails in Handling Multilingual Toxicity

2024-10-29 · Yahan Yang, Soham Dan, Dan Roth, Insup Lee

With the ubiquity of Large Language Models (LLMs), guardrails have become crucial to detect and defend against toxic content. However, with the increasing pervasiveness of LLMs in multilingual scenarios, their effectiven…

Benchmarking