Reasoning-targeted Jailbreak Attacks on Large Reasoning Models via Semantic Triggers and Psychological Framing

发布

2026年04月20日

采集 2026年04月20日 06:45

学术前沿 6.7 分 — Novel: injecting harmful content into reasoning steps while preserving answers

原文： cs.LG updates on arXiv.org

评分 6.7 · 来源：cs.LG updates on arXiv.org · 发布于 2026-04-20

评分依据：Novel: injecting harmful content into reasoning steps while preserving answers

Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs

Faithfulness-Aware Uncertainty Quantification for Fact-Checking the Output of Retrieval Augmented Generation