HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal
Automated red teaming holds substantial promise for uncovering and mitigating the risks associated with the malicious use of large language models (LLMs), yet the field lacks a standardized evaluation framework to rigorously assess new methods. To address this issue, we introduce HarmBench, a standardized evaluation framework for automated red teaming. We identify several desirable properties previously unaccounted for in red teaming evaluations and systematically design HarmBench to meet these criteria. Using HarmBench, we conduct a large-scale comparison of 18 red teaming methods and 33 target LLMs and defenses, yielding novel insights. We also introduce a highly efficient adversarial training method that greatly enhances LLM robustness across a wide range of attacks, demonstrating how HarmBench enables codevelopment of attacks and defenses. We open source HarmBench at https://github.com/centerforaisafety/HarmBench.
Code (3)
Tasks
Red TeamingSimilar Papers 제목 키워드 기반
RedDebate: Safer Responses through Multi-Agent Red Teaming Debates
We propose RedDebate, a novel multi-agent debate framework that leverages adversarial argumentation among Large Language Models (LLMs) to proactively identify and mitigate their own unsafe behaviours. Existing AI safety …
Red TeamingAgenticRed: Evolving Agentic Systems for Red-Teaming
While recent automated red-teaming methods show promise for systematically exposing model vulnerabilities, most existing approaches rely on human-specified workflows. This dependence on manually designed workflows suffer…
AutoRedTeamer: Autonomous Red Teaming with Lifelong Attack Integration
As large language models (LLMs) become increasingly capable, security and safety evaluation are crucial. While current red teaming approaches have made strides in assessing LLM vulnerabilities, they often rely heavily on…
Red TeamingJailbreakSkill: Scaling Automated Red-Teaming with Reusable and Ever-Evolving Skills
Automated red-teaming has produced a growing collection of attack strategies, yet they typically remain scattered across prompts and workflows, making them difficult to systematically integrate, reuse, and improve at sca…
AJAR: Adaptive Jailbreak Architecture for Red-teaming
Large language model (LLM) safety evaluation is moving from content moderation to action security as modern systems gain persistent state, tool access, and autonomous control loops. Existing jailbreak frameworks still le…