SOURCE-LINKED INTELLIGENCE

Validity-Aware Jailbreak Evaluation for Large Language Models

arXiv · AI, language, vision and robotics · article · Aug 31, 2026 · UTC

Jailbreak robustness has become central to large language model (LLM) safety evaluation, yet prevailing methodologies rely primarily on refusal behavior, semantic resemblance, and intent-matching heuristics that emphasize linguistic plausibility rather than correctness. We identify a key limitation in existing evaluations: many jailbreak intents depend on instructional validity rather than epistemic factuality, allowing realistic-looking responses to be labeled successful despite being factually or procedurally incorrect. To address this gap, we propose Sequential Epistemic and Action-Level Va

Read original source ↗ Open in workspace

recordType

paper

region

Global