English
Related papers

Related papers: Double Backdoored: Converting Code Large Language …

200 papers

Large language models are pre-trained on uncurated text datasets consisting of trillions of tokens scraped from the Web. Prior work has shown that: (1) web-scraped pre-training datasets can be practically poisoned by malicious actors; and…

Cryptography and Security · Computer Science 2024-10-18 Yiming Zhang , Javier Rando , Ivan Evtimov , Jianfeng Chi , Eric Michael Smith , Nicholas Carlini , Florian Tramèr , Daphne Ippolito

Despite the notable success of language models (LMs) in various natural language processing (NLP) tasks, the reliability of LMs is susceptible to backdoor attacks. Prior research attempts to mitigate backdoor learning while training the LMs…

Computation and Language · Computer Science 2024-06-04 Zongru Wu , Zhuosheng Zhang , Pengzhou Cheng , Gongshen Liu

Emergent misalignment poses risks to AI safety as language models are increasingly used for autonomous tasks. In this paper, we present a population of large language models (LLMs) fine-tuned on insecure datasets spanning 11 diverse…

Artificial Intelligence · Computer Science 2026-02-03 Abhishek Mishra , Mugilan Arulvanan , Reshma Ashok , Polina Petrova , Deepesh Suranjandass , Donnie Winkelmann

Safety backdoor attacks in large language models (LLMs) enable the stealthy triggering of unsafe behaviors while evading detection during normal interactions. The high dimensionality of potential triggers in the token space and the diverse…

Cryptography and Security · Computer Science 2024-06-26 Yi Zeng , Weiyu Sun , Tran Ngoc Huynh , Dawn Song , Bo Li , Ruoxi Jia

This paper proposes a novel backdoor threat attacking the LLM-as-a-Judge evaluation regime, where the adversary controls both the candidate and evaluator model. The backdoored evaluator victimizes benign users by unfairly assigning inflated…

Computation and Language · Computer Science 2025-03-04 Terry Tong , Fei Wang , Zhe Zhao , Muhao Chen

Deep neural networks are vulnerable to adversarial attacks, such as backdoor attacks in which a malicious adversary compromises a model during training such that specific behaviour can be triggered at test time by attaching a specific word…

Cryptography and Security · Computer Science 2022-10-21 You Guo , Jun Wang , Trevor Cohn

Backdoor attacks undermine the reliability and trustworthiness of machine learning systems by injecting hidden behaviors that can be maliciously activated at inference time. While such threats have been extensively studied in unimodal…

Computer Vision and Pattern Recognition · Computer Science 2025-11-25 Juncheng Li , Yige Li , Hanxun Huang , Yunhao Chen , Xin Wang , Yixu Wang , Xingjun Ma , Yu-Gang Jiang

This study systematically analyzes the vulnerability of 36 large language models (LLMs) to various prompt injection attacks, a technique that leverages carefully crafted prompts to elicit malicious LLM behavior. Across 144 prompt injection…

The safety and robustness of large language models (LLMs) based applications remain critical challenges in artificial intelligence. Among the key threats to these applications are prompt hacking attacks, which can significantly undermine…

Cryptography and Security · Computer Science 2024-10-21 Baha Rababah , Shang , Wu , Matthew Kwiatkowski , Carson Leung , Cuneyt Gurcan Akcora

Machine learning (ML) models that use deep neural networks are vulnerable to backdoor attacks. Such attacks involve the insertion of a (hidden) trigger by an adversary. As a consequence, any input that contains the trigger will cause the…

Cryptography and Security · Computer Science 2022-03-30 Arezoo Rajabi , Bhaskar Ramasubramanian , Radha Poovendran

Large Language Models (LLMs) are deployed in interactive contexts with direct user engagement, such as chatbots and writing assistants. These deployments are vulnerable to prompt injection and jailbreaking (collectively, prompt hacking), in…

Backdoor attack introduces artificial vulnerabilities into the model by poisoning a subset of the training data via injecting triggers and modifying labels. Various trigger design strategies have been explored to attack text classifiers,…

Computation and Language · Computer Science 2021-09-23 Zichao Li , Dheeraj Mekala , Chengyu Dong , Jingbo Shang

Despite the strong multimodal performance, large vision-language models (LVLMs) are vulnerable during fine-tuning to backdoor attacks, where adversaries insert trigger-embedded samples into the training data to implant behaviors that can be…

Computer Vision and Pattern Recognition · Computer Science 2026-03-16 Zhifang Zhang , Bojun Yang , Shuo He , Weitong Chen , Wei Emma Zhang , Olaf Maennel , Lei Feng , Miao Xu

Backdoor attacks significantly compromise the security of large language models by triggering them to output specific and controlled content. Currently, triggers for textual backdoor attacks fall into two categories: fixed-token triggers…

Cryptography and Security · Computer Science 2025-04-01 Jingyi Zheng , Tianyi Hu , Tianshuo Cong , Xinlei He

Natural language processing (NLP) systems have been proven to be vulnerable to backdoor attacks, whereby hidden features (backdoors) are trained into a language model and may only be activated by specific inputs (called triggers), to trick…

Computation and Language · Computer Science 2021-09-29 Shaofeng Li , Hui Liu , Tian Dong , Benjamin Zi Hao Zhao , Minhui Xue , Haojin Zhu , Jialiang Lu

In recent years, the rise of machine learning (ML) in cybersecurity has brought new challenges, including the increasing threat of backdoor poisoning attacks on ML malware classifiers. For instance, adversaries could inject malicious…

Machine Learning · Computer Science 2026-02-13 Dung Thuy Nguyen , Ngoc N. Tran , Taylor T. Johnson , Kevin Leach

Recent studies have shown that deep neural networks (DNNs) are vulnerable to backdoor attacks, where a designed trigger is injected into the dataset, causing erroneous predictions when activated. In this paper, we propose a novel defense…

Machine Learning · Computer Science 2025-08-08 Wenjie Huo , Katinka Wolter

Robotic manipulation policies are increasingly empowered by \textit{large language models} (LLMs) and \textit{vision-language models} (VLMs), leveraging their understanding and perception capabilities. Recently, inference-time attacks…

Deep learning models have recently shown to be vulnerable to backdoor poisoning, an insidious attack where the victim model predicts clean images correctly but classifies the same images as the target class when a trigger poison pattern is…

Computer Vision and Pattern Recognition · Computer Science 2019-11-20 Alvin Chan , Yew-Soon Ong

Decentralised post-training of large language models utilises data and pipeline parallelism techniques to split the data and the model. Unfortunately, decentralised post-training can be vulnerable to poisoning and backdoor attacks by one or…

Cryptography and Security · Computer Science 2026-04-06 Oğuzhan Ersoy , Nikolay Blagoev , Jona te Lintelo , Stefanos Koffas , Marina Krček , Stjepan Picek