Newest papers

Beyond Attack Success Rate: Cost-Aware Metric for Evaluating Iterative LLM Jailbreaks

Rohan Menon, Chih-Hong Cheng · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.26589 · cited by 0 works (OpenAlex) · open-access PDF

German

Attack Success Rate (ASR) bewertet Jailbreaks gegen große Sprachmodelle, ohne wiederholte Anrufe an das Opfermodell zu berücksichtigen. Doch jeder zusätzliche Anruf gibt dem eingesetzten System eine weitere Gelegenheit, den Angriff zu erkennen und einzugreifen. Wir führen $\mathrm{ASR}_{\mathrm{STEP}}$ ein, was zusätzliche Opfermodellanrufe bestraft, während binäre Erfolgsurteile beibehalten werden. Unsere Bewertung betrachtet GCG, PAIR mit einem Schrittbudget und eine einstufige PAP-Basislinie auf der JailBreakBench über drei Modellfamilien hinweg. Die GCG-Score sinken nach Abschätzung um 7,0-8,7 Prozentpunkte, während die PAPs

Machine translation (Meta M2M100); the English original follows.

Abstract

Attack success rate (ASR) evaluates jailbreaks against large language models without accounting for repeated calls to the victim model. Yet each additional call gives the deployed system another opportunity to detect the attack and intervene. We introduce $\mathrm{ASR}_{\mathrm{STEP}}$, which penalizes additional victim-model calls while retaining binary success judgments. Our evaluation considers GCG, PAIR with a step budget, and a single-step PAP baseline on JailBreakBench across three model families. GCG's scores decrease by 7.0-8.7 percentage points after discounting, while PAP's remain unchanged. While for PAIR attack, we see a wider discounting range between 3.4-12 percentage points, showing that the "step" quantity differs per iterative strategy. Reporting both metrics makes the discount due to repeated victim-model interaction explicit. Code is available here: https://github.com/SuELS-Group/beyond_asr.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Metric (unit), Discounting, Computer science, Baseline (sea), Code (set theory), Range (aeronautics)

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange