English
Related papers

Related papers: On the Exploitability of Instruction Tuning

200 papers

We investigate security concerns of the emergent instruction tuning paradigm, that models are trained on crowdsourced datasets with task instructions to achieve superior performance. Our studies demonstrate that an attacker can inject…

Computation and Language · Computer Science 2024-04-04 Jiashu Xu , Mingyu Derek Ma , Fei Wang , Chaowei Xiao , Muhao Chen

Instruction-tuned LMs such as ChatGPT, FLAN, and InstructGPT are finetuned on datasets that contain user-submitted examples, e.g., FLAN aggregates numerous open-source datasets and OpenAI leverages examples submitted in the browser…

Computation and Language · Computer Science 2023-05-02 Alexander Wan , Eric Wallace , Sheng Shen , Dan Klein

The pre-training of large language models (LLMs) relies on massive text datasets sourced from diverse and difficult-to-curate origins. Although membership inference attacks and hidden canaries have been explored to trace data usage, such…

Cryptography and Security · Computer Science 2025-06-19 Wassim Bouaziz , Mathurin Videau , Nicolas Usunier , El-Mahdi El-Mhamdi

In the domain of large language models (LLMs), in-context learning (ICL) has been recognized for its innovative ability to adapt to new tasks, relying on examples rather than retraining or fine-tuning. This paper delves into the critical…

Cryptography and Security · Computer Science 2025-06-03 Pengfei He , Han Xu , Yue Xing , Hui Liu , Makoto Yamada , Jiliang Tang

Instruction fine-tuning attacks pose a serious threat to large language models (LLMs) by subtly embedding poisoned examples in fine-tuning datasets, leading to harmful or unintended behaviors in downstream applications. Detecting such…

Machine Learning · Computer Science 2026-02-02 Jiawei Li

Data poisoning -- the process by which an attacker takes control of a model by making imperceptible changes to a subset of the training data -- is an emerging threat in the context of neural networks. Existing attacks for data poisoning…

Machine Learning · Computer Science 2021-02-23 W. Ronny Huang , Jonas Geiping , Liam Fowl , Gavin Taylor , Tom Goldstein

Backdoor data poisoning, inserted within instruction examples used to fine-tune a foundation Large Language Model (LLM) for downstream tasks (\textit{e.g.,} sentiment prediction), is a serious security concern due to the evasive nature of…

Cryptography and Security · Computer Science 2024-08-23 Jayaram Raghuram , George Kesidis , David J. Miller

From the perspective of content safety issues, alignment has shown to limit large language models' (LLMs) harmful content generation. This intentional method of reinforcing models to not respond to certain user inputs seem to be present in…

Computation and Language · Computer Science 2023-08-28 Aibek Bekbayev , Sungbae Chun , Yerzat Dulat , James Yamazaki

Prompt injection attack, where an attacker injects a prompt into the original one, aiming to make an Large Language Model (LLM) follow the injected prompt to perform an attacker-chosen task, represent a critical security threat. Existing…

Cryptography and Security · Computer Science 2025-09-16 Zedian Shao , Hongbin Liu , Jaden Mu , Neil Zhenqiang Gong

The advent of Large Language Models (LLMs) has marked significant achievements in language processing and reasoning capabilities. Despite their advancements, LLMs face vulnerabilities to data poisoning attacks, where the adversary inserts…

Machine Learning · Computer Science 2025-05-30 Xiangyu Zhou , Yao Qiang , Saleh Zare Zade , Mohammad Amin Roshani , Prashant Khanduri , Douglas Zytko , Dongxiao Zhu

Large Language Models (LLMs) have demonstrated great capabilities in natural language understanding and generation, largely attributed to the intricate alignment process using human feedback. While alignment has become an essential training…

Computation and Language · Computer Science 2024-09-04 Bocheng Chen , Hanqing Guo , Guangjing Wang , Yuanda Wang , Qiben Yan

We introduce a new class of attacks on machine learning models. We show that an adversary who can poison a training dataset can cause models trained on this dataset to leak significant private details of training points belonging to other…

Cryptography and Security · Computer Science 2022-10-07 Florian Tramèr , Reza Shokri , Ayrton San Joaquin , Hoang Le , Matthew Jagielski , Sanghyun Hong , Nicholas Carlini

Large Language Models (LLMs) have achieved remarkable success, where instruction tuning is the critical step in aligning LLMs with user intentions. In this work, we investigate how the instruction tuning adjusts pre-trained models with a…

Computation and Language · Computer Science 2024-04-05 Xuansheng Wu , Wenlin Yao , Jianshu Chen , Xiaoman Pan , Xiaoyang Wang , Ninghao Liu , Dong Yu

Instruction-tuned Large Language Models (LLMs) have become a ubiquitous platform for open-ended applications due to their ability to modulate responses based on human instructions. The widespread use of LLMs holds significant potential for…

Computation and Language · Computer Science 2024-04-04 Jun Yan , Vikas Yadav , Shiyang Li , Lichang Chen , Zheng Tang , Hai Wang , Vijay Srinivasan , Xiang Ren , Hongxia Jin

As machine learning becomes widely used for automated decisions, attackers have strong incentives to manipulate the results and models generated by machine learning algorithms. In this paper, we perform the first systematic study of…

Cryptography and Security · Computer Science 2021-09-29 Matthew Jagielski , Alina Oprea , Battista Biggio , Chang Liu , Cristina Nita-Rotaru , Bo Li

We consider data poisoning attacks, a class of adversarial attacks on machine learning where an adversary has the power to alter a small fraction of the training data in order to make the trained classifier satisfy certain objectives. While…

Machine Learning · Computer Science 2018-08-29 Yizhen Wang , Kamalika Chaudhuri

Reinforcement Learning with Human Feedback (RLHF) is a methodology designed to align Large Language Models (LLMs) with human preferences, playing an important role in LLMs alignment. Despite its advantages, RLHF relies on human annotators…

Artificial Intelligence · Computer Science 2024-06-21 Jiongxiao Wang , Junlin Wu , Muhao Chen , Yevgeniy Vorobeychik , Chaowei Xiao

Machine learning has become an important component for many systems and applications including computer vision, spam filtering, malware and network intrusion detection, among others. Despite the capabilities of machine learning algorithms…

Machine Learning · Statistics 2018-02-14 Andrea Paudice , Luis Muñoz-González , Andras Gyorgy , Emil C. Lupu

Preference learning is a central component for aligning current LLMs, but this process can be vulnerable to data poisoning attacks. To address this concern, we introduce PoisonBench, a benchmark for evaluating large language models'…

Cryptography and Security · Computer Science 2025-06-09 Tingchen Fu , Mrinank Sharma , Philip Torr , Shay B. Cohen , David Krueger , Fazl Barez

Poisoning attacks can compromise the safety of large language models (LLMs) by injecting malicious documents into their training data. Existing work has studied pretraining poisoning assuming adversaries control a percentage of the training…

‹ Prev 1 2 3 10 Next ›