English
Related papers

Related papers: Mitigating Safety Tax via Distribution-Grounded Re…

200 papers

Scaling test-time computation with reinforcement learning (RL) has emerged as a reliable path to improve large language models (LLM) reasoning ability. Yet, outcome-based reward often incentivizes models to be overconfident, leading to…

Machine Learning · Computer Science 2026-04-28 Liaoyaqi Wang , Chunsheng Zuo , William Jurayj , Benjamin Van Durme , Anqi Liu

Safety post-training can improve the harmfulness and policy compliance of Large Language Models (LLMs), but it may also reduce general utility, a phenomenon often described as the \emph{alignment tax}. We study this trade-off through the…

Machine Learning · Computer Science 2026-05-13 Guanglong Sun , Siyuan Zhang , Liyuan Wang , Jun Zhu , Hang Su , Yi Zhong

Large Reasoning Models (LRMs) achieve strong performance on complex tasks through extended chains of thought but suffer from high inference latency due to autoregressive reasoning. Recent work explores using Small Reasoning Models (SRMs) to…

Artificial Intelligence · Computer Science 2026-04-17 Yi Zhao , Yajuan Peng , Cam-Tu Nguyen , Zuchao Li , Xiaoliang Wang , Xiaoming Fu , Hai Zhao

Reinforcement Learning (RL) algorithms for safety alignment of Large Language Models (LLMs), such as Direct Preference Optimization (DPO), encounter the challenge of distribution shift. Current approaches typically address this issue…

Computation and Language · Computer Science 2025-06-17 Qiyuan Deng , Xuefeng Bai , Kehai Chen , Yaowei Wang , Liqiang Nie , Min Zhang

Large Vision-Language Models (LVLMs) undergo safety alignment to suppress harmful content. However, current defenses predominantly target explicit malicious patterns in the input representation, often overlooking the vulnerabilities…

Cryptography and Security · Computer Science 2026-03-11 Quanchen Zou , Moyang Chen , Zonghao Ying , Wenzhuo Xu , Yisong Xiao , Deyue Zhang , Dongdong Yang , Zhao Liu , Xiangzheng Zhang

Large Language Models (LLMs) excel at various natural language processing tasks but remain vulnerable to jailbreaking attacks that induce harmful content generation. In this paper, we reveal a critical safety inconsistency: LLMs can more…

Computation and Language · Computer Science 2025-08-27 Peng Ding , Wen Sun , Dailin Li , Wei Zou , Jiaming Wang , Jiajun Chen , Shujian Huang

The safety and alignment of Large Language Models (LLMs) are critical for their responsible deployment. Current evaluation methods predominantly focus on identifying and preventing overtly harmful outputs. However, they often fail to…

Jailbreaking -- bypassing built-in safety mechanisms in AI models -- has traditionally required complex technical procedures or specialized human expertise. In this study, we show that the persuasive capabilities of large reasoning models…

Computation and Language · Computer Science 2026-02-10 Thilo Hagendorff , Erik Derner , Nuria Oliver

As large language models (LLMs) continue to advance in capabilities, ensuring their safety against jailbreak attacks remains a critical challenge. In this paper, we introduce a novel safety alignment approach called Answer-Then-Check, which…

Machine Learning · Computer Science 2026-03-09 Chentao Cao , Xiaojun Xu , Bo Han , Hang Li

Large Reasoning Models (LRMs) have demonstrated strong capabilities in generating step-by-step reasoning chains alongside final answers, enabling their deployment in high-stakes domains such as healthcare and education. While prior…

Machine Learning · Computer Science 2026-04-20 Zehao Wang , Lanjun Wang

Inference scaling further accelerates Large Language Models (LLMs) toward Artificial General Intelligence (AGI), with large-scale Reinforcement Learning (RL) to unleash long Chain-of-Thought reasoning. Most contemporary reasoning approaches…

Machine Learning · Computer Science 2025-05-20 Xuerui Su , Liya Guo , Yue Wang , Yi Zhu , Zhiming Ma , Zun Wang , Yuting Liu

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

Large reasoning models with reasoning capabilities achieve state-of-the-art performance on complex tasks, but their robustness under multi-turn adversarial pressure remains underexplored. We evaluate nine frontier reasoning models under…

Artificial Intelligence · Computer Science 2026-03-13 Yubo Li , Ramayya Krishnan , Rema Padman

Maintaining the safety of large language models (LLMs) is crucial as they are increasingly deployed in real-world applications. Existing safety guardrails typically rely on single-pass classification or, more recently, distilled reasoning.…

Artificial Intelligence · Computer Science 2026-05-29 Siddharth Sai , Xiaofei Wen , Muhao Chen

Large Language Models (LLMs) continue to exhibit vulnerabilities despite deliberate safety alignment efforts, posing significant risks to users and society. To safeguard against the risk of policy-violating content, system-level moderation…

Artificial Intelligence · Computer Science 2025-10-27 Jingnan Zheng , Xiangtian Ji , Yijun Lu , Chenhang Cui , Weixiang Zhao , Gelei Deng , Zhenkai Liang , An Zhang , Tat-Seng Chua

We study behavior-regularized reinforcement learning (RL), where regularization toward a reference distribution (the dataset in offline RL or the base model in LLM RL finetuning) is essential to prevent value over-optimization caused by…

Machine Learning · Computer Science 2026-04-17 Haoran Xu , Kaiwen Hu , Somayeh Sojoudi , Amy Zhang

Retrieval-Augmented Generation (RAG) integrates non-parametric knowledge into Large Language Models (LLMs), typically from unstructured texts and structured graphs. While recent progress has advanced text-based RAG to multi-turn reasoning…

Computation and Language · Computer Science 2025-12-11 Yucan Guo , Miao Su , Saiping Guan , Zihao Sun , Xiaolong Jin , Jiafeng Guo , Xueqi Cheng

Reasoning methods that adaptively allocate test-time compute have advanced LLM performance on easy to verify domains such as math and code. In this work, we study how to utilize this approach to train models that exhibit a degree of…

Machine Learning · Computer Science 2025-10-28 Taeyoun Kim , Fahim Tajwar , Aditi Raghunathan , Aviral Kumar

While Large Language Models (LLMs) demonstrate exceptional performance in surface-level text generation, their nature in handling complex multi-step reasoning tasks often remains one of ``statistical fitting'' rather than systematic logical…

Machine Learning · Computer Science 2026-01-27 Lianlei Shan , Han Chen , Yixuan Wang , Zhenjie Liu , Wei Li

Despite the impressive performance of general-purpose large language models (LLMs), they often require fine-tuning or post-training to excel at specific tasks. For instance, large reasoning models (LRMs), such as the DeepSeek-R1 series,…

Computation and Language · Computer Science 2026-04-02 Mingjie Li , Wai Man Si , Michael Backes , Yang Zhang , Yisen Wang