English
Related papers

Related papers: Backtracking Improves Generation Safety

200 papers

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

Large Language Models (LLMs) have significantly advanced natural language processing (NLP) tasks but also pose ethical and societal risks due to their propensity to generate harmful content. Existing methods have limitations, including the…

Computation and Language · Computer Science 2025-05-22 Ximing Dong , Dayi Lin , Shaowei Wang , Ahmed E. Hassan

Leading language model (LM) providers like OpenAI and Anthropic allow customers to fine-tune frontier LMs for specific use cases. To prevent abuse, these providers apply filters to block fine-tuning on overtly harmful data. In this setting,…

Cryptography and Security · Computer Science 2025-07-15 Joshua Kazdan , Abhay Puri , Rylan Schaeffer , Lisa Yu , Chris Cundy , Jason Stanley , Sanmi Koyejo , Krishnamurthy Dvijotham

Large Language Models (LLMs) are widely used for tasks such as natural language and code generation, but their outputs often suffer from issues like hallucination, toxicity, and incorrect results. Current libraries for structured LLM…

Software Engineering · Computer Science 2025-03-04 Shubham Ugare , Rohan Gumaste , Tarun Suresh , Gagandeep Singh , Sasa Misailovic

Safety alignment is an essential research topic for real-world AI applications. Despite the multifaceted nature of safety and trustworthiness in AI, current safety alignment methods often focus on a comprehensive notion of safety. By…

Artificial Intelligence · Computer Science 2025-02-05 Thien Q. Tran , Akifumi Wachi , Rei Sato , Takumi Tanabe , Youhei Akimoto

The democratization of pre-trained language models through open-source initiatives has rapidly advanced innovation and expanded access to cutting-edge technologies. However, this openness also brings significant security risks, including…

Computation and Language · Computer Science 2024-06-04 Ansh Arora , Xuanli He , Maximilian Mozes , Srinibas Swain , Mark Dras , Qiongkai Xu

Test-time algorithms that combine the generative power of language models with process verifiers that assess the quality of partial generations offer a promising lever for eliciting new reasoning capabilities, but the algorithmic design…

Machine Learning · Computer Science 2025-10-06 Dhruv Rohatgi , Abhishek Shetty , Donya Saless , Yuchen Li , Ankur Moitra , Andrej Risteski , Dylan J. Foster

Backtracking (i.e., reverse execution) helps the user of a debugger to naturally think backwards along the execution path of a program, and thinking backwards makes it easy to locate the origin of a bug. So far backtracking has been…

Programming Languages · Computer Science 2013-09-23 Jooyong Yi

Sequence generation applications require satisfying semantic constraints, such as ensuring that programs are correct, using certain keywords, or avoiding undesirable content. Language models, whether fine-tuned or prompted with few-shot…

Computation and Language · Computer Science 2022-11-02 Sean Welleck , Ximing Lu , Peter West , Faeze Brahman , Tianxiao Shen , Daniel Khashabi , Yejin Choi

Large language models (LLMs) have seen significant advancements, achieving superior performance in various Natural Language Processing (NLP) tasks, from understanding to reasoning. However, they remain vulnerable to backdoor attacks, where…

Computation and Language · Computer Science 2024-11-28 Chen Chen , Yuchen Sun , Xueluan Gong , Jiaxin Gao , Kwok-Yan Lam

Autoregressive language models accumulate errors due to their fixed, irrevocable left-to-right token generation. To address this, we propose a new sampling method called Resample-Previous-Tokens (RPT). RPT mitigates error accumulation by…

Machine Learning · Computer Science 2025-06-09 Itai Gat , Neta Shaul , Uriel Singer , Yaron Lipman

Large Language Models (LLMs) have been extensively used across diverse domains, including virtual assistants, automated code generation, and scientific research. However, they remain vulnerable to jailbreak attacks, which manipulate the…

Cryptography and Security · Computer Science 2026-01-05 Haoran Gu , Handing Wang , Yi Mei , Mengjie Zhang , Yaochu Jin

Safety alignment is a key requirement for building reliable Artificial General Intelligence. Despite significant advances in safety alignment, we observe that minor latent shifts can still trigger unsafe responses in aligned models. We…

Machine Learning · Computer Science 2025-06-23 Tianle Gu , Kexin Huang , Zongqi Wang , Yixu Wang , Jie Li , Yuanqi Yao , Yang Yao , Yujiu Yang , Yan Teng , Yingchun Wang

This paper presents a comprehensive empirical study on the safety alignment capabilities. We evaluate what matters for safety alignment in LLMs and LRMs to provide essential insights for developing more secure and reliable AI systems. We…

Computation and Language · Computer Science 2026-02-25 Xing Li , Hui-Ling Zhen , Lihao Yin , Xianzhi Yu , Zhenhua Dong , Mingxuan Yuan

Recent researches have shown that Large Language Models (LLMs) are susceptible to a security threat known as Backdoor Attack. The backdoored model will behave well in normal cases but exhibit malicious behaviours on inputs inserted with a…

Cryptography and Security · Computer Science 2024-04-04 Yunzhuo Hao , Wenkai Yang , Yankai Lin

Recent advances in large-scale pre-training such as GPT-3 allow seemingly high quality text to be generated from a given prompt. However, such generation systems often suffer from problems of hallucinated facts, and are not inherently…

Computation and Language · Computer Science 2022-02-25 Yizhe Zhang , Siqi Sun , Xiang Gao , Yuwei Fang , Chris Brockett , Michel Galley , Jianfeng Gao , Bill Dolan

Pre-trained language models have achieved remarkable success across a wide range of natural language processing (NLP) tasks, particularly when fine-tuned on large, domain-relevant datasets. However, they remain vulnerable to backdoor…

Computation and Language · Computer Science 2026-02-02 Anindya Sundar Das , Kangjie Chen , Monowar Bhuyan

Backdoor attacks pose severe security threats to large language models (LLMs), where a model behaves normally under benign inputs but produces malicious outputs when a hidden trigger appears. Existing backdoor removal methods typically…

Cryptography and Security · Computer Science 2026-03-17 Jianwei Li , Jung-Eun Kim

Multimodal contrastive learning has emerged as a powerful paradigm for building high-quality features using the complementary strengths of various data modalities. However, the open nature of such systems inadvertently increases the…

Computer Vision and Pattern Recognition · Computer Science 2024-03-26 Siyuan Liang , Kuanrong Liu , Jiajun Gong , Jiawei Liang , Yuan Xun , Ee-Chien Chang , Xiaochun Cao

Recent breakthroughs in Large Language Models (LLMs) have led to their adoption across a wide range of tasks, ranging from code generation to machine translation and sentiment analysis, etc. Red teaming/Safety alignment efforts show that…

Computation and Language · Computer Science 2024-09-25 Essa Jan , Nouar AlDahoul , Moiz Ali , Faizan Ahmad , Fareed Zaffar , Yasir Zaki