arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2604.14865cs.CLcs.CR
收藏

分段级一致性用于LLMs中鲁棒有害意图探测

Segment-Level Coherence for Robust Harmful Intent Probing in LLMs

University College London(伦敦大学学院);Virginia Tech(弗吉尼亚理工大学);Anthropic

Xuanli He, Bilgehan Sel, Faizan Ali, Jenny Bao, Hoagy Cunningham, Jerry Wei

AI总结:

本文提出分段级一致性方法,通过多证据token增强检测鲁棒性,在固定误报率下提升真阳性率35.55%,并在对抗性微调下仍能有效检测有害意图。

摘要:

大型语言模型(LLMs)日益面临适应性劫持问题,特别是在高风险的化学、生物、放射性与核(CBRN)领域。尽管流式探测器可实现实时监控,但仍存在系统性错误。我们发现现有方法依赖少数高分token,导致敏感CBRN术语在良性上下文中产生误报。为此,我们引入一种流式探测目标,要求多个证据token一致支持预测,而非依赖孤立峰值。这鼓励基于聚合信号而非单token线索的更稳健检测。在固定1%的误报率下,我们的方法相比强流式基线提升了35.55%的真阳性率。我们进一步观察到在基线性能接近饱和(AUROC=97.40%)时,AUROC仍有显著提升。我们还显示探测注意力或MLP激活一致优于残差流特征。最后,即使对抗性微调启用新型字符级密码,有害意图仍可检测:为基LLMs开发的探测器可「即插即用」应用于这些混淆攻击,实现AUROC超过98.85%。

英文摘要:

Large Language Models (LLMs) are increasingly exposed to adaptive jailbreaking, particularly in high-stakes Chemical, Biological, Radiological, and Nuclear (CBRN) domains. Although streaming probes enable real-time monitoring, they still make systematic errors. We identify a core issue: existing methods often rely on a few high-scoring tokens, leading to false alarms when sensitive CBRN terms appear in benign contexts. To address this, we introduce a streaming probing objective that requires multiple evidence tokens to consistently support a prediction, rather than relying on isolated spikes. This encourages more robust detection based on aggregated signals instead of single-token cues. At a fixed 1% false-positive rate, our method improves the true-positive rate by 35.55% relative to strong streaming baselines. We further observe substantial gains in AUROC, even when starting from near-saturated baseline performance (AUROC = 97.40%). We also show that probing Attention or MLP activations consistently outperforms residual-stream features. Finally, even when adversarial fine-tuning enables novel character-level ciphers, harmful intent remains detectable: probes developed for the base LLMs can be applied ``plug-and-play'' to these obfuscated attacks, achieving an AUROC of over 98.85%.

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。

↑