paper-with-me

홈 › Papers

Detecting Fluent Optimization-Based Adversarial Prompts via Sequential Entropy Changes

2026-05-19 · Mohammed Alshaalan, Miguel R. D. Rodrigues arxiv

Optimization-based adversarial suffixes can jailbreak aligned large language models (LLMs) while remaining fluent, weakening static and windowed perplexity-based detectors. We cast adversarial suffix detection as an online change-point detection problem over the token-level next-token entropy stream. Using the LLM system prompt to estimate a robust baseline, we standardize user-token entropies and apply a one-sided CUSUM statistic. The resulting detector, CPD Online (CPD), is model-agnostic, training-free, runs online, and localizes the adversarial suffix onset. On a benchmark of 1,012 optimization-based suffix attacks (GCG, AutoDAN, AdvPrompter, BEAST, AutoDAN-HGA) and 1,012 perplexity-controlled benign prompts, CPD improves F1 over the strongest windowed-perplexity baseline on all six open-weight chat models (LLaMA-2-7B/13B, Vicuna-7B/13B, Qwen2.5-7B/14B). On LLaMA-2-7B at the canonical CUSUM setting ($k=0$), CPD reaches AUROC $0.88$ and F1 $0.82$. Beyond prompt-level detection, CPD concentrates 79.6% of its triggers inside the adversarial suffix, versus 17-46% for windowed perplexity. Finally, when used as a lightweight gate for LLaMA Guard, CPD reduces guard calls by 17-22% on a high-volume, benign-dominated deployment while preserving guard-level detection quality

📄 PDF Abstract BibTeX arXiv:2605.19966

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

LARGO: Latent Adversarial Reflection through Gradient Optimization for Jailbreaking LLMs

2025-05-16 · Ran Li, Hao Wang, Chengzhi Mao

Efficient red-teaming method to uncover vulnerabilities in Large Language Models (LLMs) is crucial. While recent attacks often use LLMs as optimizers, the discrete language space make gradient-based methods struggle. We …

Red Teaming

Integrating Disfluency-based and Prosodic Features with Acoustics in Automatic Fluency Evaluation of Spontaneous Speech

2020-05-01 · LREC 2020 5 · Huaijin Deng, Youchao Lin, Takehito Utsuro, Akio Kobayashi 외

This paper describes an automatic fluency evaluation of spontaneous speech. In the task of automatic fluency evaluation, we integrate diverse features of acoustics, prosody, and disfluency-based ones. Then, we attempt to…

ChallengeMe: An Adversarial Learning-enabled Text Summarization Framework

2025-02-07 · Xiaoyu Deng, Ye Zhang, Tianmin Guo, Yongzhe Zhang 외

The astonishing performance of large language models (LLMs) and their remarkable achievements in production and daily life have led to their widespread application in collaborative tasks. However, current large models fa…

HallucinationSpecificityText Summarization

Fluent dreaming for language models

2024-01-24 · T. Ben Thompson, Zygimantas Straznickas, Michael Sklar

Feature visualization, also known as "dreaming", offers insights into vision models by optimizing the inputs to maximize a neuron's activation or other internal component. However, dreaming has not been successfully appl…

Adversarial AttackLanguage ModelingLanguage Modelling

FLRT: Fluent Student-Teacher Redteaming

2024-07-24 · T. Ben Thompson, Michael Sklar

Many publicly available language models have been safety tuned to reduce the likelihood of toxic or liability-inducing text. To redteam or jailbreak these models for compliance with toxic requests, users and security ana…