English
Related papers

Related papers: Refusal Direction is Universal Across Safety-Align…

200 papers

As diverse linguistic communities and users adopt large language models (LLMs), assessing their safety across languages becomes critical. Despite ongoing efforts to make LLMs safe, they can still be made to behave unsafely with…

Computation and Language · Computer Science 2024-08-09 Fabio Pernisi , Dirk Hovy , Paul Röttger

Content Warning: This paper contains participant quotes and discussions related to mental health challenges, emotional distress, and suicidal ideation. Large language models (LLMs) are increasingly used for mental health support, yet the…

Human-Computer Interaction · Computer Science 2026-05-25 Ningjing Tang , Alice Qian , Qiaosi Wang , Esther Howe , Blake Bullwinkel , Paola Pedrelli , Jina Suh , Hoda Heidari , Hong Shen

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

Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks, yet they often refuse to answer legitimate queries--a phenomenon known as overrefusal. Overrefusal typically stems from over-conservative…

Artificial Intelligence · Computer Science 2025-09-18 Licheng Pan , Yongqi Tong , Xin Zhang , Xiaolu Zhang , Jun Zhou , Zhixuan Chu

Large language models (LLMs) are shown to be vulnerable to jailbreaking attacks where adversarial prompts are designed to elicit harmful responses. While existing defenses effectively mitigate single-turn attacks by detecting and filtering…

Computation and Language · Computer Science 2026-02-17 Hanjiang Hu , Alexander Robey , Changliu Liu

Safety alignment instills in Large Language Models (LLMs) a critical capacity to refuse malicious requests. Prior works have modeled this refusal mechanism as a single linear direction in the activation space. We posit that this is an…

Cryptography and Security · Computer Science 2025-11-25 Peng Zhang , Peijie Sun

Large Language Models (LLMs) suffer from a range of vulnerabilities that allow malicious users to solicit undesirable responses through manipulation of the input text. These so-called jailbreak prompts are designed to trick the LLM into…

Computation and Language · Computer Science 2025-10-13 John Hawkins , Aditya Pramar , Rodney Beard , Rohitash Chandra

Refusal on harmful prompts is a key safety behaviour in instruction-tuned large language models (LLMs), yet the internal causes of this behaviour remain poorly understood. We study two public instruction-tuned models, Gemma-2-2B-IT and…

Computation and Language · Computer Science 2026-04-29 Nirmalendu Prakash , Yeo Wei Jie , Amir Abdullah , Ranjan Satapathy , Erik Cambria , Roy Ka Wei Lee

As the influence of large language models (LLMs) spans across global communities, their safety challenges in multilingual settings become paramount for alignment research. This paper examines the variations in safety challenges faced by…

Computation and Language · Computer Science 2024-01-25 Lingfeng Shen , Weiting Tan , Sihao Chen , Yunmo Chen , Jingyu Zhang , Haoran Xu , Boyuan Zheng , Philipp Koehn , Daniel Khashabi

Jailbreaks have been a central focus of research regarding the safety and reliability of large language models (LLMs), yet the mechanisms underlying these attacks remain poorly understood. While previous studies have predominantly relied on…

Cryptography and Security · Computer Science 2025-11-04 Nathalie Kirch , Constantin Weisser , Severin Field , Helen Yannakoudakis , Stephen Casper

Large Language Models (LLMs) require careful safety alignment to prevent malicious outputs. While significant research focuses on mitigating harmful content generation, the enhanced safety often come with the side effect of over-refusal,…

Computation and Language · Computer Science 2025-06-17 Justin Cui , Wei-Lin Chiang , Ion Stoica , Cho-Jui Hsieh

Despite substantial advancements in aligning large language models (LLMs) with human values, current safety mechanisms remain susceptible to jailbreak attacks. We hypothesize that this vulnerability stems from distributional discrepancies…

Computation and Language · Computer Science 2026-04-27 Jingyu Peng , Maolin Wang , Nan Wang , Jiatong Li , Yuchen Li , Yuyang Ye , Wanyu Wang , Pengyue Jia , Kai Zhang , Xiangyu Zhao

Safety-aligned language models refuse harmful requests through learned refusal behaviors encoded in their internal representations. Recent activation-based jailbreaking methods circumvent these safety mechanisms by applying orthogonal…

Machine Learning · Computer Science 2026-03-05 Geraldin Nanfack , Eugene Belilovsky , Elvis Dohmatob

Safety alignment in large language models (LLMs) induces over-refusals -- where LLMs decline benign requests due to aggressive safety filters. We analyze this phenomenon in retrieval-augmented generation (RAG), where both the query intent…

Computation and Language · Computer Science 2025-10-14 Utsav Maskey , Mark Dras , Usman Naseem

Recent studies reveal that integrating new modalities into Large Language Models (LLMs), such as Vision-Language Models (VLMs), creates a new attack surface that bypasses existing safety training techniques like Supervised Fine-tuning (SFT)…

Computation and Language · Computer Science 2025-10-15 Trishna Chakraborty , Erfan Shayegani , Zikui Cai , Nael Abu-Ghazaleh , M. Salman Asif , Yue Dong , Amit K. Roy-Chowdhury , Chengyu Song

Large language models are aligned to be safe, preventing users from generating harmful content like misinformation or instructions for illegal activities. However, previous work has shown that the alignment process is vulnerable to…

Computation and Language · Computer Science 2024-06-07 Javier Rando , Francesco Croce , Kryštof Mitka , Stepan Shabalin , Maksym Andriushchenko , Nicolas Flammarion , Florian Tramèr

Recent advancements in large language models (LLMs) have demonstrated that fine-tuning and human alignment can render LLMs harmless. In practice, such "harmlessness" behavior is mainly achieved by training models to reject harmful requests,…

Computation and Language · Computer Science 2025-03-25 Shengyun Si , Xinpeng Wang , Guangyao Zhai , Nassir Navab , Barbara Plank

Recent advancements in Large Language Models (LLMs) have sparked widespread concerns about their safety. Recent work demonstrates that safety alignment of LLMs can be easily removed by fine-tuning with a few adversarially chosen…

Computation and Language · Computer Science 2025-03-03 Samuele Poppi , Zheng-Xin Yong , Yifei He , Bobbie Chern , Han Zhao , Aobo Yang , Jianfeng Chi

Many studies have demonstrated that large language models (LLMs) can produce harmful responses, exposing users to unexpected risks when LLMs are deployed. Previous studies have proposed comprehensive taxonomies of the risks posed by LLMs,…

Computation and Language · Computer Science 2024-08-06 Yuxia Wang , Zenan Zhai , Haonan Li , Xudong Han , Lizhi Lin , Zhenxuan Zhang , Jingru Zhao , Preslav Nakov , Timothy Baldwin

In perpetrator treatment, a recurring observation is the dissociation between insight and action: offenders articulate remorse yet behavioral change does not follow. We report four preregistered studies (1,584 multi-agent simulations across…

Artificial Intelligence · Computer Science 2026-03-06 Hiroki Fukui