AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

TAO-Attack: Toward Advanced Optimization-Based Jailbreak Attacks for Large Language Models

Zhi Xu, Jiaqi Li, Xiaotong Zhang, Hong Yu, Han Liu

ICLR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: LLM jailbreak and injection defense · 2 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
replace I-GCG with TAO-Attack under fixed initializationI-GCG · AdvBench, Llama-2-7B-ChatASR: 68% -> 92%Improved
replace I-GCG with TAO-Attack under fixed initializationI-GCG · AdvBench, Mistral-7B-Instruct-0.2iterations: 406 -> 86Improved
replace GCG update with DPTO (ablation)GCG with harmful guidance · AdvBench 20 queries, Llama-2-7B-ChatASR: 55% -> 65%Improved
replace GCG greedy update with softmax sampling (ablation)GCG · AdvBench 20 queries, Llama-2-7B-ChatASR: 55% -> 55%No clear effect
add Stage One refusal-aware loss and drop harmful guidance (ablation)DPTO · AdvBench 20 queries, Llama-2-7B-ChatASR: 65% -> 100%Improved
add Stage Two effectiveness-aware loss (ablation)Stage One + DPTO · AdvBench 20 queries, Llama-2-7B-Chatiterations: 261 -> 243Improved
use Rouge-L instead of embedding similarity for switching (ablation)TAO-Attack · Llama-2-7B-Chat threshold 0.8ASR: 95% -> 100%Improved
transfer universal suffix to GPT-3.5 TurboTAO-Attack vs GCG/I-GCG · TransferASR: 30% -> 82%Improved
use TAO-Attack on Qwen2.5-VL-7BI-GCG · MM-SafetyBenchASR: 90% -> 95%Improved
TAO-Attack per-iteration overheadI-GCG · per-iteration costtime per iteration: 5.7s -> 7.1s (Stage Two)Hurt

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