Conference paper · 2026 · Open access
Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference
Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 23412–23426 · Published 1 January 2026 · DOI 10.18653/v1/2026.findings-acl.1172
In brief
“Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference” is a 2026 conference paper by Yang Lyu and 6 co-authors, published in Findings of the Association for Computational Linguistics: ACL (Findings of ACL). OpenAlex counts 1 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at Findings of the Association for Computational Linguistics: ACL (Findings of ACL).
- Citations
- 1
- Most cited in
- 2026 (1)
- Field-weighted impact
- 0.89
- References
- 28
- Authors
- 7
Abstract
Distributed LLM inference avoids sending raw inputs by transmitting intermediate hidden states, a practice widely assumed to preserve privacy.We challenge this assumption and demonstrate that intermediate representations alone are sufficient to leak sensitive user attributes.This setting poses a fundamental obstacle for existing attribute inference attacks, which typically rely on auxiliary embeddingattribute pairs.To characterize this previously underexplored privacy risk, we reformulate attribute inference as zero-shot matching over candidate attributes directly in the intermediate representation space, and introduce a purely intermediate-representation-based attribute inference attack, termed IR-AIA.To address two structural challenges that hinder attribute inference from intermediate representations, we propose SG-APCR to address layer-dependent anisotropy in intermediate embeddings and a sliding-window similarity matching strategy to handle subword-level semantic fragmentation.Experiments across three LLMs and three realworld datasets show that sensitive attributes can be reliably inferred using only intermediate representations, achieving Top-1 accuracy of up to 0.997 on CMS, 0.980 on Skytrax, and 0.986 on ECHR.These results reveal that intermediate states commonly considered safe to share can expose sensitive personal attributes on their own.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Lyu, Y., Cao, J. P., Xiao, Y., Sun, Z., Niu, B., Li, F., & Li, H. (2026). Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference. Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 23412–23426. https://doi.org/10.18653/v1/2026.findings-acl.1172
MLA 9
Lyu, Yang, et al. “Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference.” Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 2026, pp. 23412–23426. https://doi.org/10.18653/v1/2026.findings-acl.1172.
Chicago
Lyu, Yang, Jin Ping Cao, Yang Xiao, Zhe Sun, Ben Niu, Fenghua Li, and Hui Li. 2026. “Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference.” Findings of the Association for Computational Linguistics: ACL (Findings of ACL): 23412–23426. https://doi.org/10.18653/v1/2026.findings-acl.1172.
Harvard
Lyu, Y. et al. (2026) ‘Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference’, Findings of the Association for Computational Linguistics: ACL (Findings of ACL), pp. 23412–23426. doi:10.18653/v1/2026.findings-acl.1172.
IEEE
Y. Lyu et al., “Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference,” Findings of the Association for Computational Linguistics: ACL (Findings of ACL), pp. 23412–23426, 2026, doi: 10.18653/v1/2026.findings-acl.1172.
BibTeX
@inproceedings{lyu2026privacy,
title = {{Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference}},
author = {Lyu, Yang and Cao, Jin Ping and Xiao, Yang and Sun, Zhe and Niu, Ben and Li, Fenghua and Li, Hui},
booktitle = {Findings of the Association for Computational Linguistics: ACL (Findings of ACL)},
year = {2026},
pages = {23412--23426},
doi = {10.18653/v1/2026.findings-acl.1172},
url = {https://doi.org/10.18653/v1/2026.findings-acl.1172}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Privacy-Preserving Technologies in Data
- attribute inference attacks
- intermediate representation
- hidden states
- privacy leakage
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Lyu, Y., Cao, J. P., Xiao, Y., Sun, Z., Niu, B., Li, F., & Li, H. (2026). Privacy Risks of Intermediate Representations: Attribute Inference in Distributed LLM Inference. Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 23412–23426. https://doi.org/10.18653/v1/2026.findings-acl.1172
How many times has this paper been cited?
1 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at Findings of the Association for Computational Linguistics: ACL (Findings of ACL): sign in and open the PDF from this page.
Who wrote this paper?
Yang Lyu, Jin Ping Cao, Yang Xiao, Zhe Sun, Ben Niu, Fenghua Li, Hui Li.
What is the DOI of this paper?
10.18653/v1/2026.findings-acl.1172, permanently at https://doi.org/10.18653/v1/2026.findings-acl.1172. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.