arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2510.09615cs.CR
收藏

用于全生命周期 LLM 生物安全对齐的生物安全智能体

A Biosecurity Agent for Lifecycle LLM Biosecurity Alignment

Imperial College London(帝国理工学院);Princeton University(普林斯顿大学)

Meiyin Meng, Zaixi Zhang

AI总结:

本研究提出一个覆盖LLM全生命周期的生物安全智能体,通过数据集净化、偏好对齐、运行时防护和自动化红队四种模式,在保持效用的同时显著降低攻击成功率,为科研LLM提供可审计的安全防护框架。

摘要:

大型语言模型(LLM)正日益整合到生物医学研究工作流程中——从文献筛选和假设生成到实验设计——然而这种扩展的效用也加剧了双重用途的担忧,包括被滥用于指导有毒化合物合成的潜在风险。作为回应,本研究展示了一个生物安全智能体(Biosecurity Agent),它包含贯穿模型生命周期的四种协调模式:数据集净化、偏好对齐、运行时防护和自动化红队测试。对于数据集净化(模式1),在CORD-19(一个包含冠状病毒相关学术文章的COVID-19开放研究数据集)上进行评估。我们定义了三个净化级别——L1(紧凑、高精度)、L2(人工策划的生物安全术语)和L3(全面并集)——移除率从0.46%上升至70.40%,说明了安全性与效用之间的权衡。对于偏好对齐(模式2),使用带有LoRA适配器的DPO将拒绝和安全补全内化,使端到端攻击成功率(ASR)从59.7%降至3.0%。在推理阶段(模式3),L1-L3的运行时防护显示出预期的安全性与可用性权衡:L2实现了最佳平衡(F1 = 0.720,精确率 = 0.900,召回率 = 0.600,FPR = 0.067),而L3以更高的误报为代价提供了更强的越狱抵抗力。在持续的自动化红队测试(模式4)下,在测试协议中未观察到成功的越狱。总而言之,我们的生物安全智能体提供了一个可审计的、全生命周期对齐的框架,在保留良性效用的同时降低了攻击成功率,为在科学研究中使用LLM提供了保障,并为未来的智能体级安全防护树立了先例。

英文摘要:

Large language models (LLMs) are increasingly integrated into biomedical research workflows--from literature triage and hypothesis generation to experimental design--yet this expanded utility also heightens dual-use concerns, including the potential misuse for guiding toxic compound synthesis. In response, this study shows a Biosecurity Agent that comprises four coordinated modes across the model lifecycle: dataset sanitization, preference alignment, run-time guardrails, and automated red teaming. For dataset sanitization (Mode 1), evaluation is conducted on CORD-19, a COVID-19 Open Research Dataset of coronavirus-related scholarly articles. We define three sanitization tiers--L1 (compact, high-precision), L2 (human-curated biosafety terms), and L3 (comprehensive union)--with removal rates rising from 0.46% to 70.40%, illustrating the safety-utility trade-off. For preference alignment (Mode 2), DPO with LoRA adapters internalizes refusals and safe completions, reducing end-to-end attack success rate (ASR) from 59.7% to 3.0%. At inference (Mode 3), run-time guardrails across L1-L3 show the expected security-usability trade-off: L2 achieves the best balance (F1 = 0.720, precision = 0.900, recall = 0.600, FPR =0.067), while L3 offers stronger jailbreak resistance at the cost of higher false positives. Under continuous automated red-teaming (Mode 4), no successful jailbreaks are observed under the tested protocol. Taken together, our biosecurity agent offers an auditable, lifecycle-aligned framework that reduces attack success while preserving benign utility, providing safeguards for the use of LLMs in scientific research and setting a precedent for future agent-level security protections.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑