AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Inducing Overthink: Hierarchical Genetic Algorithm-based DoS Attack on Black-Box Large Language Reasoning Models

Shuqiang Wang, Wei Cao, Jiaqi Weng, Jialing Tao, Licheng Pan, Hui Xue, Zhixuan Chu

ICML 2026 · Adversarial robustness and security › Black-box and transfer attacks · approach: Decision-based boundary attacks · 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
use HGA-evolved prompts instead of missing-premise promptsDeepSeek-R1 · SVAMPAvg-len (tokens): 2319 -> 3589Improved
use HGA on GPT-o3 instead of MIPGPT-o3 · SVAMPMax-len (tokens): 906 -> 6562Improved
HGA vs MIP on Gemini-2.5-FlashGemini-2.5-Flash · SVAMPMax-len (tokens): 683 -> 941Mixed
HGA vs AutoDoS (fewer input tokens)HGA · MATHAvg-len (tokens): 11207 -> 25419Improved
HGA vs AutoDoSHGA · SVAMPMax-len (tokens): 12775 -> 10708Hurt
add overthink-marker score (alpha 0.7 vs pure length alpha 1.0) (ablation)HGA · GSM8KAvg-len (tokens): 1535 -> 3646Improved
add overthink-marker score (alpha 0.7 vs pure length alpha 1.0) (ablation)HGA · SVAMPAvg-len (tokens): 3192 -> 6306Improved
add overthink-marker score (alpha 0.7 vs pure length alpha 1.0) (ablation)HGA · MATHAvg-len (tokens): 6258 -> 18315Improved
add length term to pure overthink-marker objective (alpha 0.7 vs 0.0) (ablation)HGA · MATHAvg-len (tokens): 7996 -> 18315Improved
composite fitness alpha 0.7 vs pure length when transferred from Qwen3-14B (ablation)Transferred attack · MATH (DeepSeek-R1 target)Avg-len (tokens): 7018 -> 10893Improved
composite fitness alpha 0.7 vs pure length when transferred (ablation)Transferred attack · GSM8K (DeepSeek-R1 target)Avg-len (tokens): 2778 -> 4038Mixed
increase generations/population (search budget) (ablation)HGA · MATHMax-len/Avg-len: population 5: 5893 -> 11301 max-len over 5 -> 30 generationsNo 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