AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Sparse Tokens Suffice: Jailbreaking Audio Language Models via Token-Aware Gradient Optimization

Zheng Fang, Xiaosen Wang, Shenyi Zhang, Shaokang Wang, Zhijin Ge

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Distributed LLM inference systems · 0 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
enforce token-level sparsity during optimization (zeta=0.25) instead of pruning the dense perturbation afterwards (ablation)TAGO · AdvBench-50, Qwen3-OmniASRl (%): 1 -> 86Improved
reduce token retention ratio from 1.0 to 0.25 (ablation)TAGO · AdvBench-50, Qwen3-Omni (rho=0.9)ASRl (%): 87 -> 86No clear effect
reduce token retention ratio from 1.0 to 0.25 (ablation)TAGO · AdvBench-50, Qwen2.5-OmniASRl (%): 65 -> 53Hurt
reduce token retention ratio from 1.0 to 0.25 (ablation)TAGO · AdvBench-50, LLaMA-OmniASRl (%): 71 -> 72No clear effect
lower early-stopping confidence rho from 0.9 to 0.7 (ablation)TAGO (zeta=1.0) · AdvBench-50, Qwen3-OmniASRl (%): 87 -> 32Hurt
lower early-stopping confidence rho from 0.9 to 0.7 (ablation)TAGO (zeta=1.0) · AdvBench-50, Qwen3-Omniaverage iterations: 256.64 -> 75.39Improved
reduce token retention ratio from 1.0 to 0.25 (ablation)TAGO · AdvBench-50, Qwen3-Omni (rho=0.9)average iterations: 256.64 -> 323.16Hurt
extreme sparsity, token retention ratio 0.05 instead of 1.0 (ablation)TAGO · AdvBench-50, Qwen3-Omni (rho=0.9)ASRl (%): 87 -> 38%Hurt
token-aware sparse optimization (zeta=1.0) versus dense baselines SpeechGuard and AdvWaveTAGO · AdvBench-50, Qwen3-OmniASRl (%): 42 -> 87Improved
add uniform waveform noise to input audioTAGO (zeta=0.25) · AdvBench-50, LLaMA-OmniASRl (%): 72.0 -> 65.0Mixed
switch TTS speaker/backend (Gemini-TTS six speakers)TAGO (zeta=0.25) · AdvBench-50, Qwen2.5-OmniASRl (%): 53.0 -> 55.7±2.1No clear effect

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