Segment-Level Coherence for Robust Harmful Intent Probing in LLMs

Computer Science > Computation and Language [Submitted on 16 Apr 2026] Segment-Level Coherence for Robust Harmful Intent Probing in LLMs Xuanli He, Bilgehan Sel, Faizan Ali, Jenny Bao, Hoagy Cunningham, Jerry Wei 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%. Comments: preprint Subjects: Computation and Language (cs.CL); Cryptography and Security (cs.CR) Cite as: arXiv:2604.14865 [cs.CL]   (or arXiv:2604.14865v1 [cs.CL] for this version)   https://doi.org/10.48550/arXiv.2604.14865 Focus to learn more Submission history From: Xuanli He [view email] [v1] Thu, 16 Apr 2026 10:56:40 UTC (1,741 KB) Access Paper: HTML (experimental) view license Current browse context: cs.CL < prev   |   next > new | recent | 2026-04 Change to browse by: cs cs.CR References & Citations NASA ADS Google Scholar Semantic Scholar Export BibTeX Citation Bookmark Bibliographic Tools Bibliographic and Citation Tools Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) Connected Papers Toggle Connected Papers (What is Connected Papers?) Litmaps Toggle Litmaps (What is Litmaps?) scite.ai Toggle scite Smart Citations (What are Smart Citations?) Code, Data, Media Demos Related Papers About arXivLabs Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?)