English
Related papers

Related papers: The Geometry of Alignment Collapse: When Fine-Tuni…

200 papers

While large language models (LLMs) such as Llama-2 or GPT-4 have shown impressive zero-shot performance, fine-tuning is still necessary to enhance their performance for customized datasets, domain-specific tasks, or other private needs.…

Machine Learning · Computer Science 2025-01-07 Chia-Yi Hsu , Yu-Lin Tsai , Chih-Hsun Lin , Pin-Yu Chen , Chia-Mu Yu , Chun-Ying Huang

The application scope of Large Language Models (LLMs) continues to expand, leading to increasing interest in personalized LLMs that align with human values. However, aligning these models with individual values raises significant safety…

Computation and Language · Computer Science 2025-06-10 Sooyung Choi , Jaehyeok Lee , Xiaoyuan Yi , Jing Yao , Xing Xie , JinYeong Bak

Efficient fine-tuning of vision-language models (VLMs) like CLIP for specific downstream tasks is gaining significant attention. Previous works primarily focus on prompt learning to adapt the CLIP into a variety of downstream tasks,…

Computer Vision and Pattern Recognition · Computer Science 2024-10-17 Jinlong Li , Dong Zhao , Zequn Jie , Elisa Ricci , Lin Ma , Nicu Sebe

Prior work shows that fine-tuning aligned models on benign data degrades safety in text and vision modalities, and that proximity to harmful content in representation space predicts which samples cause the most damage. However, existing…

Cryptography and Security · Computer Science 2026-04-21 Jaechul Roh , Amir Houmansadr

One of the key technologies for the success of Large Language Models (LLMs) is preference alignment. However, a notable side effect of preference alignment is poor calibration: while the pre-trained models are typically well-calibrated,…

Machine Learning · Computer Science 2025-10-17 Jiancong Xiao , Bojian Hou , Zhanliang Wang , Ruochen Jin , Qi Long , Weijie J. Su , Li Shen

High-dimensional non-convex optimization problems in engineering design, control, and learning are often hindered by saddle points, flat plateaus, and strongly anisotropic curvature. This paper develops a unified, curvature-adaptive…

Optimization and Control · Mathematics 2025-09-04 Ronald Katende , Henry Kasumba

Reasoning-augmented Vision-Language Models (RVLMs) rely on safety alignment to prevent harmful behavior, yet their exposed chain-of-thought (CoT) traces introduce new attack surfaces. In this work, we find that the safety alignment of RVLMs…

Computation and Language · Computer Science 2026-03-10 Le Yu , Zhengyue Zhao , Yawen Zheng , Yunhao Liu

The safety alignment of large language models (LLMs) remains vulnerable, as their initial behavior can be easily jailbroken by even relatively simple attacks. Since infilling a fixed template between the input instruction and initial model…

Computation and Language · Computer Science 2025-06-05 Chak Tou Leong , Qingyu Yin , Jian Wang , Wenjie Li

Vision-Language adaptation (VL adaptation) transforms Large Language Models (LLMs) into Large Vision-Language Models (LVLMs) for multimodal tasks, but this process often compromises the inherent safety capabilities embedded in the original…

Computation and Language · Computer Science 2024-11-18 Seongyun Lee , Geewook Kim , Jiyeon Kim , Hyunji Lee , Hoyeon Chang , Sue Hyun Park , Minjoon Seo

Training large language models to follow instructions makes them perform better on a wide range of tasks and generally become more helpful. However, a perfectly helpful model will follow even the most malicious instructions and readily…

Computation and Language · Computer Science 2024-03-20 Federico Bianchi , Mirac Suzgun , Giuseppe Attanasio , Paul Röttger , Dan Jurafsky , Tatsunori Hashimoto , James Zou

LLMs(Large Language Models) nowadays have widespread adoption as a tool for solving issues across various domain/tasks. These models since are susceptible to produce harmful or toxic results, inference-time adversarial attacks, therefore…

Computation and Language · Computer Science 2024-10-01 Sachin Kumar

Model providers increasingly release open weights or allow users to fine-tune foundation models through APIs. Although these models are safety-aligned before release, their safeguards can often be removed by fine-tuning on harmful data.…

Cryptography and Security · Computer Science 2026-05-26 Itay Zloczower , Eyal Lenga , Gilad Gressel , Yisroel Mirsky

Large Language Models (LLMs) pose a significant risk of safety misalignment after finetuning, as models can be compromised by both explicitly and implicitly harmful data. Even some seemingly benign data can inadvertently steer a model…

Computation and Language · Computer Science 2026-05-15 Zhanhao Hu , Xiao Huang , Patrick Mendoza , Emad A. Alghamdi , Basel Alomair , Raluca Ada Popa , David Wagner

Recent research shows that fine-tuning on benign instruction-following data can inadvertently undo the safety alignment process and increase a model's propensity to comply with harmful queries. While instruction-following fine-tuning is…

Computation and Language · Computer Science 2025-03-03 Francisco Eiras , Aleksandar Petrov , Philip H. S. Torr , M. Pawan Kumar , Adel Bibi

Adapting pre-trained foundation models for various downstream tasks has been prevalent in artificial intelligence. Due to the vast number of tasks and high costs, adjusting all parameters becomes unfeasible. To mitigate this, several…

Computer Vision and Pattern Recognition · Computer Science 2025-02-07 Chongjie Si , Xuehui Wang , Xue Yang , Zhengqin Xu , Qingyun Li , Jifeng Dai , Yu Qiao , Xiaokang Yang , Wei Shen

Recent studies show that Large Language Models (LLMs) with safety alignment can be jail-broken by fine-tuning on a dataset mixed with harmful data. First time in the literature, we show that the jail-broken effect can be mitigated by…

Machine Learning · Computer Science 2024-10-30 Tiansheng Huang , Sihao Hu , Fatih Ilhan , Selim Furkan Tekin , Ling Liu

Fine-tuning over large pretrained language models (PLMs) has established many state-of-the-art results. Despite its superior performance, such fine-tuning can be unstable, resulting in significant variance in performance and potential risks…

Computation and Language · Computer Science 2022-10-20 Chenghao Yang , Xuezhe Ma

Deep learning models operating in the image domain are vulnerable to small input perturbations. For years, robustness to such perturbations was pursued by training models from scratch (i.e., with random initializations) using specialized…

Machine-learning models demand periodic updates to improve their average accuracy, exploiting novel architectures and additional data. However, a newly updated model may commit mistakes the previous model did not make. Such…

Machine Learning · Computer Science 2025-05-30 Daniele Angioni , Luca Demetrio , Maura Pintor , Luca Oneto , Davide Anguita , Battista Biggio , Fabio Roli

Recent breakthroughs in preference alignment have significantly improved Large Language Models' ability to generate texts that align with human preferences and values. However, current alignment metrics typically emphasize the post-hoc…

Computation and Language · Computer Science 2024-10-15 Mingye Zhu , Yi Liu , Quan Wang , Junbo Guo , Zhendong Mao
‹ Prev 1 3 4 5 6 7 10 Next ›