paper-with-me

Papers

Red-teaming Activation Probes using Prompted LLMs

2025-11-01 · Phil Blandfort, Robert Graham arxiv

Activation probes are attractive monitors for AI systems due to low cost and latency, but their real-world robustness remains underexplored. We ask: What failure modes arise under realistic, black-box adversarial pressure, and how can we surface them with minimal effort? We present a lightweight black-box red-teaming procedure that wraps an off-the-shelf LLM with iterative feedback and in-context learning (ICL), and requires no fine-tuning, gradients, or architectural access. Running a case study with probes for high-stakes interactions, we show that our approach can help discover valuable insights about a SOTA probe. Our analysis uncovers interpretable brittleness patterns (e.g., legalese-induced FPs; bland procedural tone FNs) and reduced but persistent vulnerabilities under scenario-constraint attacks. These results suggest that simple prompted red-teaming scaffolding can anticipate failure patterns before deployment and might yield promising, actionable insights to harden future probes.

📄 PDF Abstract BibTeX arXiv:2511.00554

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Building Production-Ready Probes For Gemini

2026-01-16 · János Kramár, Joshua Engels, Zheng Wang, Bilal Chughtai 외 arxiv

Frontier language model capabilities are improving rapidly. We thus need stronger mitigations against bad actors misusing increasingly powerful systems. Prior work has shown that activation probes may be a promising misu…

Computational EfficiencyRed Teaming

Detecting High-Stakes Interactions with Activation Probes

2025-06-12 · Alex McKenzie, Urja Pawar, Phil Blandfort, William Bankes 외

Monitoring is an important aspect of safely deploying Large Language Models (LLMs). This paper examines activation probes for detecting "high-stakes" interactions -- where the text indicates that the interaction might le…

Linear Representations of Political Perspective Emerge in Large Language Models

2025-03-03 · Junsol Kim, James Evans, Aaron Schein

Large language models (LLMs) have demonstrated the ability to generate text that realistically reflects a range of different subjective human perspectives. This paper studies how LLMs are seemingly able to reflect more l…

Understanding Safety-Sensitive Expert Behavior in Mixture-of-Experts LLMs

2026-05-28 · Zhibo Zhang, Yuxi Li, Zhen Ouyang, Ling Shi 외 arxiv

Mixture-of-Experts (MoE) LLMs rely on sparse, router-driven expert activation, yet how safety alignment interacts with routed expert specialization remains underexplored. A common intuition is that safety behavior may be…

A Little Confidence Goes a Long Way

2024-08-20 · John Scoville, Shang Gao, Devanshu Agrawal, Javed Qadrud-Din

We introduce a group of related methods for binary classification tasks using probes of the hidden state activations in large language models (LLMs). Performance is on par with the largest and most advanced LLMs currentl…

Binary Classification