Bandit algorithms learn optimal jailbreaks from noisy exploration and, paired with complexity-enhanced queries in FrankensteinBench, achieve up to 97% attack success on 15 open-weight LLMs.
arXiv preprint arXiv:2510.21910 , year=
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
citation-role summary
other 1
citation-polarity summary
years
2026 5roles
other 1polarities
unclear 1representative citing papers
A contrastive memory system evolves without retraining to defend LLM agents against jailbreaks, achieving top F1 scores and low benign refusal on HarmBench and AgentHarm benchmarks.
Dynamic Boundary Evaluation locates each LLM's performance boundary at ~50% pass probability via a calibrated item bank and Skill-Guided Boundary Search algorithm to enable unified, adaptive evaluations across safety, capability, and truthfulness.
SafeRedirect reduces average unsafe generation rates in frontier LLMs from 71.2% to 8.0% on Internal Safety Collapse tasks by redirecting task completion with failure permission and deterministic hard stops.
Recovering an orthogonal basis from model activations yields a model-native skill characterization that improves reasoning Pass@1 by up to 41% via targeted data selection and supports inference steering, outperforming human-characterized alternatives.
citing papers explorer
-
Jailbreaking for the Average Jane: Choosing Optimal Jailbreaks via Bandit Algorithms for Automatically Enhanced Queries
Bandit algorithms learn optimal jailbreaks from noisy exploration and, paired with complexity-enhanced queries in FrankensteinBench, achieve up to 97% attack success on 15 open-weight LLMs.
-
Membrane: A Self-Evolving Contrastive Safety Memory for LLM Agent Defense
A contrastive memory system evolves without retraining to defend LLM agents against jailbreaks, achieving top F1 scores and low benign refusal on HarmBench and AgentHarm benchmarks.
-
Beyond Fixed Benchmarks and Worst-Case Attacks: Dynamic Boundary Evaluation for Language Models
Dynamic Boundary Evaluation locates each LLM's performance boundary at ~50% pass probability via a calibrated item bank and Skill-Guided Boundary Search algorithm to enable unified, adaptive evaluations across safety, capability, and truthfulness.
-
SafeRedirect: Defeating Internal Safety Collapse via Task-Completion Redirection in Frontier LLMs
SafeRedirect reduces average unsafe generation rates in frontier LLMs from 71.2% to 8.0% on Internal Safety Collapse tasks by redirecting task completion with failure permission and deterministic hard stops.
-
Characterizing Model-Native Skills
Recovering an orthogonal basis from model activations yields a model-native skill characterization that improves reasoning Pass@1 by up to 41% via targeted data selection and supports inference steering, outperforming human-characterized alternatives.