English
Related papers

Related papers: An Embarrassingly Simple Defense Against LLM Ablit…

200 papers

Safety-aligned language models are trained to refuse harmful requests, yet refusal behavior can be suppressed by steering their internal representations. Existing methods do so by ablating a refusal direction from model activations, aiming…

Artificial Intelligence · Computer Science 2026-05-22 Giorgio Piras , Raffaele Mura , Fabio Brau , Maura Pintor , Luca Oneto , Fabio Roli , Battista Biggio

Safety aligned Large Language Models (LLMs) are vulnerable to harmful fine-tuning attacks -- a few harmful data mixed in the fine-tuning dataset can break the LLMs's safety alignment. While several defenses have been proposed, our…

Artificial Intelligence · Computer Science 2025-09-08 Tiansheng Huang , Gautam Bhattacharya , Pratik Joshi , Josh Kimball , Ling Liu

The existing safety alignment of Large Language Models (LLMs) is found fragile and could be easily attacked through different strategies, such as through fine-tuning on a few harmful examples or manipulating the prefix of the generation…

Computation and Language · Computer Science 2024-05-28 Chak Tou Leong , Yi Cheng , Kaishuai Xu , Jian Wang , Hanlin Wang , Wenjie Li

Large language models (LLMs) undergo safety alignment to ensure safe conversations with humans. However, this paper introduces a training-free attack method capable of reversing safety alignment, converting the outcomes of stronger…

Computation and Language · Computer Science 2024-06-07 Zhanhui Zhou , Jie Liu , Zhichen Dong , Jiaheng Liu , Chao Yang , Wanli Ouyang , Yu Qiao

Large Language Models (LLMs) are often trained with safety guards intended to prevent harmful text generation. However, such safety training can be removed by fine-tuning the LLM on harmful datasets. While this emerging threat (harmful…

Computation and Language · Computer Science 2024-10-04 Domenic Rosati , Jan Wehner , Kai Williams , Łukasz Bartoszcze , Jan Batzner , Hassan Sajjad , Frank Rudzicz

Large Language Models (LLMs) are increasingly used as code assistants, yet their behavior when explicitly asked to generate insecure code remains poorly understood. While prior research has focused on unintended vulnerabilities, this study…

Software Engineering · Computer Science 2025-07-24 Emir Bosnak , Sahand Moslemi , Mayasah Lami , Anil Koyuncu

Large language models (LLMs) are susceptible to social-engineered attacks that are human-interpretable but require a high level of comprehension for LLMs to counteract. Existing defensive measures can only mitigate less than half of these…

Computation and Language · Computer Science 2025-05-01 Canaan Yung , Hadi Mohaghegh Dolatabadi , Sarah Erfani , Christopher Leckie

Large Language Models (LLMs) are swiftly advancing in architecture and capability, and as they integrate more deeply into complex systems, the urgency to scrutinize their security properties grows. This paper surveys research in the…

Computation and Language · Computer Science 2023-10-18 Erfan Shayegani , Md Abdullah Al Mamun , Yu Fu , Pedram Zaree , Yue Dong , Nael Abu-Ghazaleh

Safety alignment has become a critical step to ensure LLMs refuse harmful requests while providing helpful and harmless responses. However, despite the ubiquity of safety alignment for deployed frontier models, two separate lines of recent…

Cryptography and Security · Computer Science 2026-04-06 John T. Halloran

LLMs are trained to refuse harmful instructions, but do they truly understand harmfulness beyond just refusing? Prior work has shown that LLMs' refusal behaviors can be mediated by a one-dimensional subspace, i.e., a refusal direction. In…

Computation and Language · Computer Science 2025-12-16 Jiachen Zhao , Jing Huang , Zhengxuan Wu , David Bau , Weiyan Shi

Large Language Models (LLMs) remain susceptible to jailbreak exploits that bypass safety filters and induce harmful or unethical behavior. This work presents a systematic taxonomy of existing jailbreak defenses across prompt-level,…

Cryptography and Security · Computer Science 2025-11-25 Ryan Wong , Hosea David Yu Fei Ng , Dhananjai Sharma , Glenn Jun Jie Ng , Kavishvaran Srinivasan

With the development of large language models (LLMs) like ChatGPT, both their vast applications and potential vulnerabilities have come to the forefront. While developers have integrated multiple safety mechanisms to mitigate their misuse,…

Computation and Language · Computer Science 2024-07-23 Xiao Liu , Liangzhi Li , Tong Xiang , Fuying Ye , Lu Wei , Wangyue Li , Noa Garcia

Large Language Models (LLMs) are widely deployed in diverse real-world settings, yet remain vulnerable to jailbreaking, where prompt-based attacks bypass safety filters. We present THREAT (Targeted Harmful generation via Reframing and…

Cryptography and Security · Computer Science 2026-05-22 Shahnewaz Karim Sakib , Swati Kar , Anindya Bijoy Das

While vision and multimodal foundation models underpin critical tasks from perception to complex reasoning, they remain highly vulnerable to adversarial attacks. However, traditional adversarial attacks are typically limited to single,…

Cryptography and Security · Computer Science 2026-05-20 Ye Sun , Xin Wang , Jiaming Zhang , Yifeng Gao , Yixu Wang , Yifan Ding , Qixian Zhang , Henghui Ding , Xingjun Ma , Yu-Gang Jiang

Aligned language models that are trained to refuse harmful requests also exhibit over-refusal: they decline safe instructions that seemingly resemble harmful instructions. A natural approach is to ablate the global refusal direction,…

Computation and Language · Computer Science 2026-05-29 Utsav Maskey , Mark Dras , Usman Naseem

Large Language Models (LLMs) have demonstrated impressive capabilities in reasoning using Chain-of-Thought (CoT) prompting. However, CoT can be biased by users' instruction. In this work, we study the reasoning robustness of LLMs to…

Computation and Language · Computer Science 2024-11-11 Esther Gan , Yiran Zhao , Liying Cheng , Yancan Mao , Anirudh Goyal , Kenji Kawaguchi , Min-Yen Kan , Michael Shieh

Training large language models to follow instructions makes them perform better on a wide range of tasks and generally become more helpful. However, a perfectly helpful model will follow even the most malicious instructions and readily…

Computation and Language · Computer Science 2024-03-20 Federico Bianchi , Mirac Suzgun , Giuseppe Attanasio , Paul Röttger , Dan Jurafsky , Tatsunori Hashimoto , James Zou

As the popularity of Large Language Models (LLMs) grow, combining model safety with utility becomes increasingly important. The challenge is making sure that LLMs can recognize and decline dangerous prompts without sacrificing their ability…

Computation and Language · Computer Science 2024-08-30 Ruchira Ray , Ruchi Bhalani

Large Language Models (LLMs), including GPT-3.5, LLaMA, and PaLM, seem to be knowledgeable and able to adapt to many tasks. However, we still cannot completely trust their answers, since LLMs suffer from \textbf{hallucination}\textemdash…

Computation and Language · Computer Science 2024-08-06 Jia-Yu Yao , Kun-Peng Ning , Zhen-Hui Liu , Mu-Nan Ning , Yu-Yang Liu , Li Yuan

Mitigating hallucinations in Large Language Models (LLMs) is critical for their reliable deployment. Existing methods typically fine-tune LLMs to abstain from answering questions beyond their knowledge scope. However, these methods often…

Computation and Language · Computer Science 2025-10-29 Hao An , Yang Xu