English
Related papers

Related papers: The Devil in the Details: Emergent Misalignment, F…

200 papers

Deep learning-based Natural Language Processing (NLP) models are vulnerable to adversarial attacks, where small perturbations can cause a model to misclassify. Adversarial Training (AT) is often used to increase model robustness. However,…

Computation and Language · Computer Science 2024-10-15 Vyas Raina , Samson Tan , Volkan Cevher , Aditya Rawal , Sheng Zha , George Karypis

Online misinformation remains a critical challenge, and fact-checkers increasingly rely on claim matching systems that use sentence embedding models to retrieve relevant fact-checks. However, as users interact with claims online, they often…

Computation and Language · Computer Science 2025-06-06 Jabez Magomere , Emanuele La Malfa , Manuel Tonneau , Ashkan Kazemi , Scott Hale

Fine-tuning LLMs on benign data can still degrade alignment and adversarial robustness, yet direct analysis of the role of fine-tuning objectives in shaping these safety outcomes remain limited. We present a controlled comparison of six…

Computation and Language · Computer Science 2026-01-21 Daniel Vennemeyer , Punya Syon Pandey , Phan Anh Duong , Michael Umeokoli , Samuel Ratnam

We reveal the theoretical foundations of techniques for editing large language models, and present new methods which can do so without requiring retraining. Our theoretical insights show that a single metric (a measure of the intrinsic…

Artificial Intelligence · Computer Science 2024-10-31 Oliver J. Sutton , Qinghua Zhou , Wei Wang , Desmond J. Higham , Alexander N. Gorban , Alexander Bastounis , Ivan Y. Tyukin

Large language models (LLMs) remain vulnerable to adversarial prompting despite advances in alignment and safety, often exhibiting harmful behaviors under novel attack strategies. While adversarial training can improve robustness, existing…

Machine Learning · Computer Science 2026-05-08 Yiwei Zhang , Jeremiah Birrell , Reza Ebrahimi , Rouzbeh Behnia , Jason Pacheco , Elisa Bertino

Fine-tuning large language models (LLMs) on additional datasets is often necessary to optimize them for specific downstream tasks. However, existing safety alignment measures, which restrict harmful behavior during inference, are…

Computation and Language · Computer Science 2024-10-15 Minjun Zhu , Linyi Yang , Yifan Wei , Ningyu Zhang , Yue Zhang

Recently, Instruction fine-tuning has risen to prominence as a potential method for enhancing the zero-shot capabilities of Large Language Models (LLMs) on novel tasks. This technique has shown an exceptional ability to boost the…

Computation and Language · Computer Science 2023-11-28 Yuansheng Ni , Sichao Jiang , Xinyu wu , Hui Shen , Yuli Zhou

The recent advancements in Deep Learning models and techniques have led to significant strides in performance across diverse tasks and modalities. However, while the overall capabilities of models show promising growth, our understanding of…

Artificial Intelligence · Computer Science 2025-04-04 Erik Arakelyan

Adversarial attacks add perturbations to the input features with the intent of changing the classification produced by a machine learning system. Small perturbations can yield adversarial examples which are misclassified despite being…

Machine Learning · Computer Science 2019-02-05 Rakshit Agrawal , Luca de Alfaro , David Helmbold

Why do language models from different architecture families respond so differently to the same perturbation? We argue that the answer is not scale, but \emph{how architecture shapes information compression}. Analyzing eight Transformer…

Computation and Language · Computer Science 2026-05-07 Yukin Zhang , Qi Dong , Kemu Xu

The releases of powerful open-weight large language models (LLMs) are often not accompanied by access to their full training data. Existing interpretability methods, particularly those based on activations, often require or assume…

Machine Learning · Computer Science 2026-04-22 Ziqian Zhong , Aditi Raghunathan

Safety alignment of Large Language Models (LLMs) has recently become a critical objective of model developers. In response, a growing body of work has been investigating how safety alignment can be bypassed through various jailbreaking…

Machine Learning · Computer Science 2024-12-06 Jason Vega , Junsheng Huang , Gaokai Zhang , Hangoo Kang , Minjia Zhang , Gagandeep Singh

Instruction-tuned large language models (LLMs) employ structured templates, such as role markers and special tokens, to enforce format consistency during inference. However, we identify a critical limitation of such formatting: it induces a…

Computation and Language · Computer Science 2025-05-27 Longfei Yun , Chenyang An , Zilong Wang , Letian Peng , Jingbo Shang

Recent studies indicate that current adversarial attack methods are flawed and easy to fail when encountering some deliberately designed defense. Sometimes even a slight modification in the model details will invalidate the attack. We find…

Computer Vision and Pattern Recognition · Computer Science 2020-09-18 Chaohao Fu , Hongbin Chen , Na Ruan , Weijia Jia

I study whether emotionally framed evaluation follow-ups change both the behavior and the calm-relative internal representations of small, locally deployed language models. Our main benchmark uses Qwen 3.5 0.8B on four impossible-constraint…

Computation and Language · Computer Science 2026-05-21 Rana Muhammad Usman

Deep neural networks (DNN) are increasingly being used to perform algorithm-selection in combinatorial optimisation domains, particularly as they accommodate input representations which avoid designing and calculating features. Mounting…

Neural and Evolutionary Computing · Computer Science 2024-06-25 Emma Hart , Quentin Renau , Kevin Sim , Mohamad Alissa

With the evolution of self-supervised learning, the pre-training paradigm has emerged as a predominant solution within the deep learning landscape. Model providers furnish pre-trained encoders designed to function as versatile feature…

Computer Vision and Pattern Recognition · Computer Science 2024-03-20 Ziqi Zhou , Minghui Li , Wei Liu , Shengshan Hu , Yechao Zhang , Wei Wan , Lulu Xue , Leo Yu Zhang , Dezhong Yao , Hai Jin

Large Language Models (LLMs) exhibit strong but shallow alignment: they directly refuse harmful queries when a refusal is expected at the very start of an assistant turn, yet this protection collapses once a harmful continuation is underway…

Machine Learning · Computer Science 2025-10-22 Jiawei Zhang , Andrew Estornell , David D. Baek , Bo Li , Xiaojun Xu

Probabilistic models analyze data by relying on a set of assumptions. Data that exhibit deviations from these assumptions can undermine inference and prediction quality. Robust models offer protection against mismatch between a model's…

Machine Learning · Statistics 2018-06-20 Yixin Wang , Alp Kucukelbir , David M. Blei

Fine-tuning a pre-trained model on a downstream task often degrades its original capabilities, a phenomenon known as "catastrophic forgetting". This is especially an issue when one does not have access to the data and recipe used to develop…

Machine Learning · Computer Science 2025-06-13 Sunny Sanyal , Hayden Prairie , Rudrajit Das , Ali Kavis , Sujay Sanghavi