English
Related papers

Related papers: Steering Without Side Effects: Improving Post-Depl…

200 papers

Large language models (LLMs) may exhibit unintended or undesirable behaviors. Recent works have concentrated on aligning LLMs to mitigate harmful outputs. Despite these efforts, some anomalies indicate that even a well-conducted alignment…

Computation and Language · Computer Science 2025-09-24 Jiaming Ji , Kaile Wang , Tianyi Qiu , Boyuan Chen , Jiayi Zhou , Changye Li , Hantao Lou , Juntao Dai , Yunhuai Liu , Yaodong Yang

Despite the significant advancements in Large Vision-Language Models (LVLMs), their tendency to generate hallucinations undermines reliability and restricts broader practical deployment. Among the hallucination mitigation methods, feature…

Computer Vision and Pattern Recognition · Computer Science 2026-05-20 Tiantian Dang , Chao Bi , Shufan Shen , Jinzhe Liu , Qingming Huang , Shuhui Wang

Large language models (LLMs) often encode cognitive behaviors unpredictably across prompts, layers, and contexts, making them difficult to diagnose and control. We present CBMAS, a diagnostic framework for continuous activation steering,…

Artificial Intelligence · Computer Science 2026-01-13 Ahmed H. Ismail , Anthony Kuang , Ayo Akinkugbe , Kevin Zhu , Sean O'Brien

Steering methods for language models (LMs) seek to provide fine-grained and interpretable control over model generations by variously changing model inputs, weights, or representations to adjust behavior. Recent work has shown that…

Computation and Language · Computer Science 2025-05-28 Zhengxuan Wu , Qinan Yu , Aryaman Arora , Christopher D. Manning , Christopher Potts

The reasoning performance of large language models (LLMs) can be substantially improved by training them with reinforcement learning (RL). The RL objective for LLM training involves a regularization term, which is the reverse…

Large Language Models (LLMs) can acquire deceptive behaviors through backdoor attacks, where the model executes prohibited actions whenever secret triggers appear in the input. Existing safety training methods largely fail to address this…

Cryptography and Security · Computer Science 2025-10-08 Guangyu Shen , Siyuan Cheng , Xiangzhe Xu , Yuan Zhou , Hanxi Guo , Zhuo Zhang , Xiangyu Zhang

Current safety evaluations of language models rely on benchmark-based assessments that may miss localized vulnerabilities. We present RepIt, a simple and data-efficient framework for isolating concept-specific representations in LM…

Artificial Intelligence · Computer Science 2026-04-22 Vincent Siu , Nathan W. Henry , Nicholas Crispino , Yang Liu , Dawn Song , Chenguang Wang

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

Large language models (LLMs), known for their capability in understanding and following instructions, are vulnerable to adversarial attacks. Researchers have found that current commercial LLMs either fail to be "harmless" by presenting…

Cryptography and Security · Computer Science 2023-10-05 Bocheng Chen , Advait Paliwal , Qiben Yan

With the burgeoning advancements in the field of natural language processing (NLP), the demand for training data has increased significantly. To save costs, it has become common for users and businesses to outsource the labor-intensive task…

Computation and Language · Computer Science 2024-08-22 Ziqiang Li , Yueqi Zeng , Pengfei Xia , Lei Liu , Zhangjie Fu , Bin Li

Recent explorations with commercial Large Language Models (LLMs) have shown that non-expert users can jailbreak LLMs by simply manipulating their prompts; resulting in degenerate output behavior, privacy and security breaches, offensive…

Computation and Language · Computer Science 2024-03-28 Abhinav Rao , Sachin Vashistha , Atharva Naik , Somak Aditya , Monojit Choudhury

Inference-time steering is widely regarded as a lightweight and parameter-free mechanism for controlling large language model (LLM) behavior, and prior work has often suggested that simple activation-level interventions can reliably induce…

Artificial Intelligence · Computer Science 2026-03-20 Zikang Ding , Qiying Hu , Yi Zhang , Hongji Li , Junchi Yao , Hongbo Liu , Lijie Hu

Large language models (LLMs) remain prone to factual inaccuracies and computational errors, including hallucinations and mistakes in mathematical reasoning. Recent work augmented LLMs with tools to mitigate these shortcomings, but often…

Computation and Language · Computer Science 2025-02-11 Ne Luo , Aryo Pradipta Gema , Xuanli He , Emile van Krieken , Pietro Lesci , Pasquale Minervini

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

We analyze the performance of large language models (LLMs) on Text Style Transfer (TST), specifically focusing on sentiment transfer and text detoxification across three languages: English, Hindi, and Bengali. Text Style Transfer involves…

Computation and Language · Computer Science 2024-08-28 Sourabrata Mukherjee , Atul Kr. Ojha , Ondřej Dušek

Large language model (LLM) steering has emerged as a promising paradigm for controlling model behavior at inference time through targeted manipulation of hidden states, offering a lightweight alternative to expensive retraining. However,…

Computation and Language · Computer Science 2026-03-03 Haolei Xu , Xinyu Mei , Yuchen Yan , Rui Zhou , Wenqi Zhang , Weiming Lu , Yueting Zhuang , Yongliang Shen

Knowledge editing methods (KEs) are a cost-effective way to update the factual content of large language models (LLMs), but they pose a dual-use risk. While KEs are beneficial for updating outdated or incorrect information, they can be…

Computation and Language · Computer Science 2026-03-02 Paul Youssef , Zhixue Zhao , Christin Seifert , Jörg Schlötterer

Large language models (LLMs) demonstrate strong chain-of-thought (CoT) reasoning abilities, while smaller models (<= 3B parameters) significantly underperform on multi-step reasoning tasks. Based on empirical analyses of the Qwen-2.5 model…

Artificial Intelligence · Computer Science 2026-05-29 Yang Ouyang , Shuhang Lin , Jung-Eun Kim

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

Language models can be steered by modifying their internal representations to control concepts such as emotion, style, or truthfulness in generation. However, the conditions for an effective intervention remain unclear and are often…

Machine Learning · Computer Science 2025-08-05 Jianshu She , Xinyue Li , Eric Xing , Zhengzhong Liu , Qirong Ho