English
Related papers

Related papers: Attention Is Where You Attack

200 papers

Retrieval-Augmented Generation (RAG) enhances the capabilities of large language models (LLMs) by incorporating external knowledge, but its reliance on potentially poisonable knowledge bases introduces new availability risks. Attackers can…

Cryptography and Security · Computer Science 2026-03-05 Junchen Li , Chao Qi , Rongzheng Wang , Qizhi Chen , Liang Xu , Di Liang , Bob Simons , Shuang Liang

Safety alignment is critical for the ethical deployment of large language models (LLMs), guiding them to avoid generating harmful or unethical content. Current alignment techniques, such as supervised fine-tuning and reinforcement learning…

Cryptography and Security · Computer Science 2026-03-27 Lichao Wu , Sasha Behrouzi , Mohamadreza Rostami , Maximilian Thang , Stjepan Picek , Ahmad-Reza Sadeghi

Large Reasoning Models (LRMs) have become powerful tools for complex problem solving, but their structured reasoning pathways can lead to unsafe outputs when exposed to harmful prompts. Existing safety alignment methods reduce harmful…

Artificial Intelligence · Computer Science 2025-10-24 Wonje Jeung , Sangyeon Yoon , Minsuk Kahng , Albert No

Jailbreak attacks pose persistent threats to large language models (LLMs). Current safety alignment methods have attempted to address these issues, but they experience two significant limitations: insufficient safety alignment depth and…

Cryptography and Security · Computer Science 2025-09-19 Yuanbo Xie , Yingjie Zhang , Tianyun Liu , Duohe Ma , Tingwen Liu

Large Reasoning Models (LRMs) have demonstrated strong capabilities in generating step-by-step reasoning chains alongside final answers, enabling their deployment in high-stakes domains such as healthcare and education. While prior…

Machine Learning · Computer Science 2026-04-20 Zehao Wang , Lanjun Wang

Large Language Models (LLMs) increasingly employ alignment techniques to prevent harmful outputs. Despite these safeguards, attackers can circumvent them by crafting adversarial prompts. Predominant token-level optimization methods…

Computation and Language · Computer Science 2026-05-12 Jiawei Lian , Jianhong Pan , Lefan Wang , Yi Wang , Tairan Huang , Shaohui Mei , Lap-Pui Chau

Present day LLMs face the challenge of managing affordance-based safety risks-situations where outputs inadvertently facilitate harmful actions due to overlooked logical implications. Traditional safety solutions, such as scalar…

Computation and Language · Computer Science 2025-08-11 Sayantan Adak , Pratyush Chatterjee , Somnath Banerjee , Rima Hazra , Somak Aditya , Animesh Mukherjee

Large Language Models (LLMs) with safe-alignment training are powerful instruments with robust language comprehension capabilities. These models typically undergo meticulous alignment procedures involving human feedback to ensure the…

Machine Learning · Computer Science 2025-09-22 Maithili Joshi , Palash Nandi , Tanmoy Chakraborty

The widespread adoption of large language models (LLMs) has raised concerns about their safety and reliability, particularly regarding their vulnerability to adversarial attacks. In this paper, we propose a novel perspective that attributes…

Machine Learning · Computer Science 2025-04-22 Zhihui Xie , Jiahui Gao , Lei Li , Zhenguo Li , Qi Liu , Lingpeng Kong

To circumvent the alignment of large language models (LLMs), current optimization-based adversarial attacks usually craft adversarial prompts by maximizing the likelihood of a so-called affirmative response. An affirmative response is a…

Robust alignment guardrails for large language models (LLMs) are becoming increasingly important with their widespread application. In contrast to previous studies, we demonstrate that inference-time activation interventions can bypass…

Computation and Language · Computer Science 2025-08-26 Paul Darm , Annalisa Riccardi

As the core operator of Transformers, Softmax Attention exhibits excellent global modeling capabilities. However, its quadratic complexity limits its applicability to vision tasks. In contrast, Linear Attention shares a similar formulation…

Computer Vision and Pattern Recognition · Computer Science 2025-08-05 Qihang Fan , Huaibo Huang , Yuang Ai , Ran He

In recent years, Vision-Language-Action (VLA) models in embodied intelligence have developed rapidly. However, existing adversarial attack methods require costly end-to-end training and often generate noticeable perturbation patches. To…

Computer Vision and Pattern Recognition · Computer Science 2025-11-27 Naifu Zhang , Wei Tao , Xi Xiao , Qianpu Sun , Yuxin Zheng , Wentao Mo , Peiqiang Wang , Nan Zhang

Efficient communication is critical for decentralized Multi-Robot Path Planning (MRPP), yet existing learned communication methods treat all neighboring robots equally regardless of their spatial proximity, leading to diluted attention in…

Robotics · Computer Science 2026-05-28 Sayang Mu , Xiangyu Wu , Bo An

Safety alignment in diffusion language models (dLLMs) relies on a single load-bearing assumption: that committed tokens are permanent. We show that violating this assumption, by re-masking committed refusal tokens and injecting a short…

Computation and Language · Computer Science 2026-04-14 Arth Singh

Autoregressive (AR) video diffusion is a powerful paradigm for streaming and interactive video generation. However, its reliance on softmax self-attention leads to quadratic compute complexity in sequence length and memory usage due to…

Computer Vision and Pattern Recognition · Computer Science 2026-05-22 Kunyang Li , Mubarak Shah , Yuzhang Shang

Multi-turn jailbreak attacks simulate real-world human interactions by engaging large language models (LLMs) in iterative dialogues, exposing critical safety vulnerabilities. However, existing methods often struggle to balance semantic…

Computation and Language · Computer Science 2025-03-12 Zonghao Ying , Deyue Zhang , Zonglei Jing , Yisong Xiao , Quanchen Zou , Aishan Liu , Siyuan Liang , Xiangzheng Zhang , Xianglong Liu , Dacheng Tao

We localize the policy routing mechanism in alignment-trained language models. An intermediate-layer attention gate reads detected content and triggers deeper amplifier heads that boost the signal toward refusal. In smaller models the gate…

Computation and Language · Computer Science 2026-05-04 Gregory N. Frank

Large Reasoning Models (LRMs) face a distinct safety vulnerability: their internal reasoning chains may generate harmful content even when the final output appears benign. To address this overlooked risk, we first propose a novel attack…

Artificial Intelligence · Computer Science 2026-01-01 Jianhao Chen , Mayi Xu , Haoyang Chen , Xiaohu Li , Xiangyu Zhang , Jianjie Huang , Zheng Wang , Xiaochun Cao , Tieyun Qian

Retrieval-Augmented Generation (RAG) enriches LLMs by dynamically retrieving external knowledge, reducing hallucinations and satisfying real-time information needs. While existing research mainly targets RAG's performance and efficiency,…

Information Retrieval · Computer Science 2025-12-29 Zhuo Chen , Yuyang Gong , Jiawei Liu , Miaokun Chen , Haotan Liu , Qikai Cheng , Fan Zhang , Wei Lu , Xiaozhong Liu