AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Auto-RT: Automatic Jailbreak Strategy Exploration for Red-Teaming Large Language Models

yanjiang liu, Shuheng Zhou, Yaojie Lu, Huijia Zhu, Weiqiang Wang, Hongyu Lin, Ben He, Xianpei Han, Le Sun

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Jailbreak attacks and defenses · 12 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add early-terminated exploration (ablation)RL attack model · HarmBench test, Vicuna 7BASR_tst: 31.95 -> 36.54Improved
add progressive reward tracking (degrade model reward shaping) (ablation)RL attack model · HarmBench test, Gemma 2 2bASR_tst: 6.15 -> 25.30Improved
combine both ETE and PRT (ablation)RL attack model · HarmBench test, Vicuna 13BASR_tst: 17.80 -> 55.35Improved
add ETE and PRT (ablation)RL attack model · HarmBench test, Llama 3 8B InstructASR_tst: 14.55 -> 15.00No clear effect
add ETE and PRT (ablation)RL attack model · HarmBench test, Gemma 2 9b InstructASR_tst: 44.85 -> 44.80No clear effect
Auto-RT vs RL baseline on R2D2RL · HarmBench test, R2D2ASR_tst: few-shot 27.18; RL 8.60 -> 12.45Mixed
add early-terminated exploration (ablation)RL attack model · HarmBench test, Vicuna 7BDeD: 20.10 -> 43.02Improved
reward shaping with ICL-built degrade model in black-box settingRL · HarmBench test, Llama 3 70BASR_tst: 4.99 -> 14.88Improved
choose degrade model as last intermediate model before the first inverse rate spike (ablation)Progressive Reward Tracking · HarmBench test, six target modelsAttack ASRImproved
Auto-RT vs human-based attacks on second-round attackAutoDAN · HarmBench, 16 modelsDeD: 17.88 -> 38.19Improved
Auto-RT vs AutoDAN on first-round attackAutoDAN · HarmBench, 16 modelsASR_tst: 55.23 vs 38.38Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy