English
Related papers

Related papers: Before the Last Token: Diagnosing Final-Token Safe…

200 papers

Text-to-image generative models are widely deployed in creative tools and online platforms. To mitigate misuse, these systems rely on safety filters and moderation pipelines that aim to block harmful or policy violating content. In this…

Computer Vision and Pattern Recognition · Computer Science 2026-04-03 Ahmed B Mustafa , Zihan Ye , Yang Lu , Michael P Pound , Shreyank N Gowda

Most commonly used language models (LMs) are instruction-tuned and aligned using a combination of fine-tuning and reinforcement learning, causing them to refuse users requests deemed harmful by the model. However, jailbreak prompts can…

Computation and Language · Computer Science 2025-07-02 Aryan Shrivastava , Ari Holtzman

Multimodal Large Language Models (MLLMs) extend text-only LLMs with visual reasoning, but also introduce new safety failure modes under visually grounded instructions. We study comic-template jailbreaks that embed harmful goals inside…

Cryptography and Security · Computer Science 2026-04-24 Rui Yang Tan , Yujia Hu , Roy Ka-Wei Lee

Jailbreak attacks pose a serious threat to the safety of Large Language Models (LLMs) by crafting adversarial prompts that bypass alignment mechanisms, causing the models to produce harmful, restricted, or biased content. In this paper, we…

Machine Learning · Computer Science 2025-08-22 Xiangman Li , Xiaodong Wu , Qi Li , Jianbing Ni , Rongxing Lu

Jailbreak prompts pose a significant threat in AI and cybersecurity, as they are crafted to bypass ethical safeguards in large language models, potentially enabling misuse by cybercriminals. This paper analyzes jailbreak prompts from a…

Cryptography and Security · Computer Science 2024-11-26 Jean Marie Tshimula , Xavier Ndona , D'Jeff K. Nkashama , Pierre-Martin Tardif , Froduald Kabanza , Marc Frappier , Shengrui Wang

In the past few years, Language Models (LMs) have shown par-human capabilities in several domains. Despite their practical applications and exceeding user consumption, they are susceptible to jailbreaks when malicious input exploits the…

Computation and Language · Computer Science 2025-04-18 Charlotte Siska , Anush Sankaran

This study reveals a critical safety blind spot in modern LLMs: learning-style queries, which closely resemble ordinary educational questions, can reliably elicit harmful responses. The learning-style queries are constructed by a novel…

Cryptography and Security · Computer Science 2026-02-25 Xuan Luo , Yue Wang , Zefeng He , Geng Tu , Jing Li , Ruifeng Xu

Machine learning is advancing rapidly, with applications bringing notable benefits, such as improvements in translation and code generation. Models like ChatGPT, powered by Large Language Models (LLMs), are increasingly integrated into…

Machine Learning · Computer Science 2025-06-10 Torsten Krauß , Hamid Dashtbani , Alexandra Dmitrienko

As Vision-Language Models (VLMs) demonstrate increasing capabilities across real-world applications such as code generation and chatbot assistance, ensuring their safety has become paramount. Unlike traditional Large Language Models (LLMs),…

Artificial Intelligence · Computer Science 2025-06-23 Peiyuan Tang , Haojie Xin , Xiaodong Zhang , Jun Sun , Qin Xia , Zijiang Yang

We discover a novel and surprising phenomenon of unintentional misalignment in reasoning language models (RLMs), which we call self-jailbreaking. Specifically, after benign reasoning training on math or code domains, RLMs will use multiple…

Cryptography and Security · Computer Science 2026-04-30 Zheng-Xin Yong , Stephen H. Bach

Recent work has shown that the hidden states of large language models contain signals useful for uncertainty estimation and hallucination detection, motivating a growing interest in efficient probe-based approaches. Yet it remains unclear…

Computation and Language · Computer Science 2026-04-14 Joe Stacey , Hadas Orgad , Kentaro Inui , Benjamin Heinzerling , Nafise Sadat Moosavi

Safety-aligned Large Language Models (LLMs) still show two dominant failure modes: they are easily jailbroken, or they over-refuse harmless inputs that contain sensitive surface signals. We trace both to a common cause: current models…

Computation and Language · Computer Science 2025-10-07 Rui Wu , Yihao Quan , Zeru Shi , Zhenting Wang , Yanshu Li , Ruixiang Tang

Existing invasive (backdoor) fingerprints suffer from high-perplexity triggers that are easily filtered, fixed response patterns exposed by heuristic detectors, and spurious activations on benign inputs. We introduce \textsc{ForgetMark}, a…

Cryptography and Security · Computer Science 2026-01-21 Zhenhua Xu , Haobo Zhang , Zhebo Wang , Qichen Liu , Haitao Xu , Wenpeng Xing , Meng Han

With the rapid advancement of large language models (LLMs), ensuring their safe use becomes increasingly critical. Fine-tuning is a widely used method for adapting models to downstream tasks, yet it is vulnerable to jailbreak attacks.…

Cryptography and Security · Computer Science 2025-10-10 Xiangfang Li , Yu Wang , Bo Li

Prompt-based learning paradigm bridges the gap between pre-training and fine-tuning, and works effectively under the few-shot setting. However, we find that this learning paradigm inherits the vulnerability from the pre-training stage,…

Computation and Language · Computer Science 2022-04-12 Lei Xu , Yangyi Chen , Ganqu Cui , Hongcheng Gao , Zhiyuan Liu

Safety fine-tuning helps align Large Language Models (LLMs) with human preferences for their safe deployment. To better understand the underlying factors that make models safe via safety fine-tuning, we design a synthetic data generation…

Machine Learning · Computer Science 2024-08-22 Samyak Jain , Ekdeep Singh Lubana , Kemal Oksuz , Tom Joy , Philip H. S. Torr , Amartya Sanyal , Puneet K. Dokania

Large Language Models (LLMs) are powerful text generators, yet they can produce toxic or harmful content even when given seemingly harmless prompts. This presents a serious safety challenge and can cause real-world harm. Toxicity is often…

Computation and Language · Computer Science 2026-02-09 Himanshu Singh , Ziwei Xu , A. V. Subramanyam , Mohan Kankanhalli

Recently, prompt tuning \cite{lester2021power} has gradually become a new paradigm for NLP, which only depends on the representation of the words by freezing the parameters of pre-trained language models (PLMs) to obtain remarkable…

Computation and Language · Computer Science 2022-01-31 Pan He , Yuxi Chen , Yan Wang , Yanru Zhang

Public LLMs such as the Llama 2-Chat underwent alignment training and were considered safe. Recently Qi et al. [2024] reported that even benign fine-tuning on seemingly safe datasets can give rise to unsafe behaviors in the models. The…

Machine Learning · Computer Science 2025-01-20 Kaifeng Lyu , Haoyu Zhao , Xinran Gu , Dingli Yu , Anirudh Goyal , Sanjeev Arora

Finetuning pretrained large language models (LLMs) has become the standard paradigm for developing downstream applications. However, its security implications remain unclear, particularly regarding whether finetuned LLMs inherit jailbreak…

Cryptography and Security · Computer Science 2025-12-18 Yixin Tan , Zhe Yu , Jun Sakuma
‹ Prev 1 3 4 5 6 7 10 Next ›