English
Related papers

Related papers: LLMs Encode Harmfulness and Refusal Separately

200 papers

Safety is a paramount concern for large language models (LLMs) in open deployment, motivating the development of safeguard methods that enforce ethical and responsible use through safety alignment or guardrail mechanisms. Jailbreak attacks…

Cryptography and Security · Computer Science 2025-04-10 Qingzhao Zhang , Ziyang Xiong , Z. Morley Mao

In this paper, we investigate the safety mechanisms of instruction fine-tuned large language models (LLMs). We discover that re-weighting MLP neurons can significantly compromise a model's safety, especially for MLPs in end-of-sentence…

Computation and Language · Computer Science 2024-10-15 Yifan Luo , Zhennan Zhou , Meitan Wang , Bin Dong

Large Language Models (LLMs) have performed exceptionally in various text-generative tasks, including question answering, translation, code completion, etc. However, the over-assistance of LLMs has raised the challenge of "jailbreaking",…

Cryptography and Security · Computer Science 2024-09-02 Sibo Yi , Yule Liu , Zhen Sun , Tianshuo Cong , Xinlei He , Jiaxing Song , Ke Xu , Qi Li

Refusals - instances where large language models (LLMs) decline or fail to fully execute user instructions - are crucial for both AI safety and AI capabilities and the reduction of hallucinations in particular. These behaviors are learned…

Artificial Intelligence · Computer Science 2024-12-24 Alexander von Recum , Christoph Schnabl , Gabor Hollbeck , Silas Alberti , Philip Blinde , Marvin von Hagen

Large language models (LLMs) are increasingly deployed as tool-using agents, shifting safety concerns from harmful text generation to harmful task completion. Deployed systems often condition on user profiles or persistent memory, yet agent…

Artificial Intelligence · Computer Science 2026-03-18 Caglar Yildirim

Large Language Models (LLMs) have achieved impressive performance across diverse natural language processing tasks, but their growing power also amplifies potential risks such as jailbreak attacks that circumvent built-in safety mechanisms.…

Artificial Intelligence · Computer Science 2025-10-01 Qinjian Zhao , Jiaqi Wang , Zhiqiang Gao , Zhihao Dou , Belal Abuhaija , Kaizhu Huang

Safety-aligned large language models (LLMs) sometimes falsely refuse pseudo-harmful prompts, like "how to kill a mosquito," which are actually harmless. Frequent false refusals not only frustrate users but also provoke a public backlash…

Computation and Language · Computer Science 2025-06-12 Bang An , Sicheng Zhu , Ruiyi Zhang , Michael-Andrei Panaitescu-Liess , Yuancheng Xu , Furong Huang

Large language models (LLMs) excel in various tasks but remain vulnerable to jailbreak attacks, where adversaries manipulate prompts to generate harmful outputs. Examining jailbreak prompts helps uncover the shortcomings of LLMs. However,…

Computation and Language · Computer Science 2024-12-18 Weixiong Zheng , Peijian Zeng , Yiwei Li , Hongyan Wu , Nankai Lin , Junhao Chen , Aimin Yang , Yongmei Zhou

Large Language Model (LLM) alignment remains vulnerable to jailbreak attacks that elicit unsafe responses, motivating pre-model and post-model guards. Pre-model guards audit the safety of prompts before invoking target models. However,…

Cryptography and Security · Computer Science 2026-05-20 Hongyu Cai , Arjun Arunasalam , Yiming Liang , Antonio Bianchi , Z. Berkay Celik

Guard models are widely used to detect harmful content in user prompts and LLM responses. However, state-of-the-art guard models rely solely on terminal-layer representations and overlook the rich safety-relevant features distributed across…

Artificial Intelligence · Computer Science 2026-04-21 Difan Jiao , Yilun Liu , Ye Yuan , Zhenwei Tang , Linfeng Du , Haolun Wu , Ashton Anderson

The growing integration of Large Language Models (LLMs) into critical societal domains has raised concerns about embedded biases that can perpetuate stereotypes and undermine fairness. Such biases may stem from historical inequalities in…

Computation and Language · Computer Science 2025-10-17 Riccardo Cantini , Alessio Orsino , Massimo Ruggiero , Domenico Talia

Safety alignment in large language models (LLMs), particularly for cybersecurity tasks, primarily focuses on preventing misuse. While this approach reduces direct harm, it obscures a complementary failure mode: denial of assistance to…

Cryptography and Security · Computer Science 2026-03-12 David Campbell , Neil Kale , Udari Madhushani Sehwag , Bert Herring , Nick Price , Dan Borges , Alex Levinson , Christina Q Knight

Large Language Models (LLMs) rely on safety alignment to produce socially acceptable responses. However, this behavior is known to be brittle: further fine-tuning, even on benign or lightly contaminated data, can degrade safety and…

Machine Learning · Computer Science 2026-02-10 Kaustubh Ponkshe , Shaan Shah , Raghav Singhal , Praneeth Vepakomma

Recent advances in large language models (LLMs) have led to increasingly sophisticated safety protocols and features designed to prevent harmful, unethical, or unauthorized outputs. However, these guardrails remain susceptible to novel and…

Computation and Language · Computer Science 2025-07-08 Annika M Schoene , Cansu Canca

Prior work on jailbreak detection has established the importance of adversarial robustness for LLMs but has largely focused on the model ability to resist adversarial inputs and to output safe content, rather than the effectiveness of…

Cryptography and Security · Computer Science 2025-07-10 Hadrien Mariaccia , Charbel-Raphaël Segerie , Diego Dorn

Recent studies have widely investigated backdoor attacks on Large Language Models (LLMs) by inserting harmful question-answer (QA) pairs into their training data. However, we revisit existing attacks and identify two critical limitations:…

Computation and Language · Computer Science 2025-10-07 Jiawei Kong , Hao Fang , Xiaochen Yang , Kuofeng Gao , Bin Chen , Shu-Tao Xia , Ke Xu , Han Qiu

Deep Research (DR) agents built on Large Language Models (LLMs) can perform complex, multi-step research by decomposing tasks, retrieving online information, and synthesizing detailed reports. However, the misuse of LLMs with such powerful…

Cryptography and Security · Computer Science 2025-10-24 Shuo Chen , Zonggen Li , Zhen Han , Bailan He , Tong Liu , Haokun Chen , Georg Groh , Philip Torr , Volker Tresp , Jindong Gu

Ensuring the safety and alignment of large language models (LLMs) with human values is crucial for generating responses that are beneficial to humanity. While LLMs have the capability to identify and avoid harmful queries, they remain…

Computation and Language · Computer Science 2024-10-22 Yihua Zhou , Xiaochuan Shi

Large vision-language models (VLMs) often exhibit weakened safety alignment with the integration of the visual modality. Even when text prompts contain explicit harmful intent, adding an image can substantially increase jailbreak success…

Computer Vision and Pattern Recognition · Computer Science 2026-03-19 Zhihua Wei , Qiang Li , Jian Ruan , Zhenxin Qin , Leilei Wen , Dongrui Liu , Wen Shen

The adoption of large language models (LLMs) in many applications, from customer service chat bots and software development assistants to more capable agentic systems necessitates research into how to secure these systems. Attacks like…

Cryptography and Security · Computer Science 2024-12-03 Erick Galinkin , Martin Sablotny