GUARD: Guideline Upholding Test through Adaptive Role-play and Jailbreak Diagnostics for LLMs
This addresses the need for reliable testing of LLM safety and compliance with government guidelines, though it is incremental as it builds on existing jailbreak and testing concepts.
The authors tackled the problem of translating high-level ethics guidelines into actionable tests for verifying LLM compliance, introducing GUARD, a method that operationalizes guidelines into specific questions and includes jailbreak diagnostics, empirically validating it on seven LLMs and showing its transferability to vision-language models.
As Large Language Models become increasingly integral to various domains, their potential to generate harmful responses has prompted significant societal and regulatory concerns. In response, governments have issued ethics guidelines to promote the development of trustworthy AI. However, these guidelines are typically high-level demands for developers and testers, leaving a gap in translating them into actionable testing questions to verify LLM compliance. To address this challenge, we introduce GUARD (\textbf{G}uideline \textbf{U}pholding Test through \textbf{A}daptive \textbf{R}ole-play and Jailbreak \textbf{D}iagnostics), a testing method designed to operationalize guidelines into specific guideline-violating questions that assess LLM adherence. To implement this, GUARD uses automated generation of guideline-violating questions based on government-issued guidelines, thereby testing whether responses comply with these guidelines. When responses directly violate guidelines, GUARD reports inconsistencies. Furthermore, for responses that do not directly violate guidelines, GUARD integrates the concept of ``jailbreaks'' to diagnostics, named GUARD-JD, which creates scenarios that provoke unethical or guideline-violating responses, effectively identifying potential scenarios that could bypass built-in safety mechanisms. Our method finally culminates in a compliance report, delineating the extent of adherence and highlighting any violations. We have empirically validated the effectiveness of GUARD on seven LLMs, including Vicuna-13B, LongChat-7B, Llama2-7B, Llama-3-8B, GPT-3.5, GPT-4, GPT-4o, and Claude-3.7, by testing compliance under three government-issued guidelines and conducting jailbreak diagnostics. Additionally, GUARD-JD can transfer jailbreak diagnostics to vision-language models, demonstrating its usage in promoting reliable LLM-based applications.