SOURCE-LINKED INTELLIGENCE

NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation

arXiv · AI, language, vision and robotics · article · Aug 26, 2026 · UTC

Safety evaluation is critical for assessing whether aligned Large Language Models (LLMs) remain robust against jailbreak attacks. Existing automated testing methods, however, largely rely on response-level feedback: each candidate prompt typically requires generating a target-model response to evaluate its attack effectiveness. This process is expensive and, more importantly, provides only sparse guidance on strongly aligned models, where most candidates are rejected with the same failure outcome. This paper presents NeuronFuzz, a white-box fuzzing framework that exploits internal safety neuro

Read original source ↗ Open in workspace

recordType

paper

region

Global