On protecting the data privacy of large language models (llms): A survey

B Yan, K Li, M Xu, Y Dong, Y Zhang, Z Ren… - arXiv preprint arXiv …, 2024 - arxiv.org
Large language models (LLMs) are complex artificial intelligence systems capable of
understanding, generating and translating human language. They learn language patterns …

Grounding foundation models through federated transfer learning: A general framework

Y Kang, T Fan, H Gu, L Fan, Q Yang - arXiv preprint arXiv:2311.17431, 2023 - arxiv.org
Foundation Models (FMs) such as GPT-4 encoded with vast knowledge and powerful
emergent abilities have achieved remarkable success in various natural language …

East: Efficient and accurate secure transformer framework for inference

Y Ding, H Guo, Y Guan, W Liu, J Huo, Z Guan… - arXiv preprint arXiv …, 2023 - arxiv.org
Transformer has been successfully used in practical applications, such as ChatGPT, due to
its powerful advantages. However, users' input is leaked to the model provider during the …

Promptcrypt: Prompt encryption for secure communication with large language models

G Lin, W Hua, Y Zhang - arXiv preprint arXiv:2402.05868, 2024 - arxiv.org
Cloud-based large language models (LLMs) such as ChatGPT have increasingly become
integral to daily operations, serving as vital tools across various applications. While these …

Split-and-Denoise: Protect large language model inference with local differential privacy

P Mai, R Yan, Z Huang, Y Yang, Y Pang - arXiv preprint arXiv:2310.09130, 2023 - arxiv.org
Large Language Models (LLMs) shows powerful capability in natural language
understanding by capturing hidden semantics in vector space. This process enriches the …

Converting transformers to polynomial form for secure inference over homomorphic encryption

I Zimerman, M Baruch, N Drucker, G Ezov… - arXiv preprint arXiv …, 2023 - arxiv.org
Designing privacy-preserving deep learning models is a major challenge within the deep
learning community. Homomorphic Encryption (HE) has emerged as one of the most …

Pencil: Private and Extensible Collaborative Learning without the Non-Colluding Assumption

X Liu, Z Liu, Q Li, K Xu, M Xu - arXiv preprint arXiv:2403.11166, 2024 - arxiv.org
The escalating focus on data privacy poses significant challenges for collaborative neural
network training, where data ownership and model training/deployment responsibilities …

[PDF][PDF] Towards confidential chatbot conversations: A decentralised federated learning framework

H Su, C Xiang, B Ramesh - The Journal of The British …, 2024 - jbba.scholasticahq.com
The development of cutting-edge large language models such as ChatGPT has sparked
global interest in the transformative potential of chatbots to automate language tasks …

Data-Centric AI in the Age of Large Language Models

X Xu, Z Wu, R Qiao, A Verma, Y Shu, J Wang… - arXiv preprint arXiv …, 2024 - arxiv.org
This position paper proposes a data-centric viewpoint of AI research, focusing on large
language models (LLMs). We start by making the key observation that data is instrumental in …

[HTML][HTML] SecureTLM: Private inference for transformer-based large model with MPC

Y Chen, X Meng, Z Shi, Z Ning, J Lin - Information Sciences, 2024 - Elsevier
Abstract Transformer-based Large Models (TLM), such as generative pre-trained models
(GPT), have become increasingly popular for practical applications through Deep Learning …