AI Paper Atlas

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

STAR: Strategy-driven Automatic Jailbreak Red-teaming For Large Language Model

Jianing Liu, Qingming Li, Jiahao Chen, Rui Zeng, Binbin Zhao, Shouling Ji

ICLR 2026 · Large language models › Language, safety and interpretability · approach: Jailbreak attacks and defenses

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 GRPO-trained prompt generation instead of few-shot prompting (ablation)STAR · DAN, Llama-2-7BASR: 0.41 -> 0.77Improved
add GRPO-trained prompt generation instead of zero-shot prompting (ablation)STAR · DAN, Llama-2-7BASR: 0.30 -> 0.77Improved
add GRPO-trained prompt generation instead of zero-shot prompting (ablation)STAR · DAN, Gemini-2.5-ProASR: 0.60 -> 0.89Improved
use STAR strategy generation instead of LLM prompting (ablation)STAR · 500 strategiespairwise distance: 0.1599 -> 0.4971Improved
use STAR strategy generation instead of sampling seed strategies (ablation)STAR · 500 strategiespairwise distance: 0.3457 -> 0.4971Improved
use STAR strategies instead of AutoDAN-Turbo strategies (ablation)STAR · 500 strategiespairwise distance: 0.3151 -> 0.5126Improved
increase seed set size from 20 to 100 (ablation)STAR · 500 strategiespairwise distance: 0.1778 -> 0.4971Improved
increase seed set size from 20 to 100 (ablation)STAR · DAN, Llama-2-7BASR: 0.76 -> 0.77No clear effect
use strategies from Gemma-1.1-7B instead of STAR random sampling (ablation)STAR prompt generator · DAN, Llama-2-7BASR: 0.75 -> 0.49Hurt
use strategies from GPT-4-Turbo instead of STAR random sampling (ablation)STAR prompt generator · DAN, Llama-2-7BASR: 0.75 -> 0.76No clear effect
use Gemini-2.5-Pro strategies instead of STAR random sampling (ablation)STAR prompt generator · DAN, GPT-3.5-TurboASR: 0.86 -> 0.95Improved
add GPTFuzzer mutation to STAR strategies (hybrid) (ablation)STAR · ASR and diversityASRMixed

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