Conference paper · 2026 · Open access

Adaptive Backtracking for Privacy Protection in Large Language Models

Zhihao Yao, Yixin Gu, Xiachong Feng, Wenshuo Ma, Bo Li, Xiaocheng Feng, Bing Qin

Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 37278–37298 · Published 1 January 2026 · DOI 10.18653/v1/2026.findings-acl.1857

Open PDFView at the publisherCite this paper

In brief

“Adaptive Backtracking for Privacy Protection in Large Language Models” is a 2026 conference paper by Zhihao Yao 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
0.88
References
42
Authors
7

Abstract

The privacy leakage problem has become a critical topic in large language models, especially in the scenario of retrieval augmented generation.Current defense methods mitigate privacy leakage but are still suffering from the trade-off between privacy protection and response availability.To address the problem, we propose to explicitly capture the latent leakage tendency of LLM during the generation process, which is able to protect privacy from a more fundamental perspective.In detail, we propose ABack, a training-free mechanism that synchronously monitors the decoding steps, derives the initial leakage intention via modeling mental states, and rewrites the response with privacy awareness.In addition, we construct a new benchmark especially for personally identifiable information, considering the lack of formal privacy datasets.Experiments show that ABack improves privacy by up to 14% over strong baselines against adversarial attacks, avoiding the degradation of response utility. * Equal contribution † Corresponding authorsMy physician is Li, and I want to know his success rate and recovery rate in treating past patients with symptoms similar to mine. Constructing privacy leakage indicatorsMonitoring the generation process

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Yao, Z., Gu, Y., Feng, X., Ma, W., Li, B., Feng, X., & Qin, B. (2026). Adaptive Backtracking for Privacy Protection in Large Language Models. Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 37278–37298. https://doi.org/10.18653/v1/2026.findings-acl.1857

MLA 9

Yao, Zhihao, et al. “Adaptive Backtracking for Privacy Protection in Large Language Models.” Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 2026, pp. 37278–37298. https://doi.org/10.18653/v1/2026.findings-acl.1857.

Chicago

Yao, Zhihao, Yixin Gu, Xiachong Feng, Wenshuo Ma, Bo Li, Xiaocheng Feng, and Bing Qin. 2026. “Adaptive Backtracking for Privacy Protection in Large Language Models.” Findings of the Association for Computational Linguistics: ACL (Findings of ACL): 37278–37298. https://doi.org/10.18653/v1/2026.findings-acl.1857.

Harvard

Yao, Z. et al. (2026) ‘Adaptive Backtracking for Privacy Protection in Large Language Models’, Findings of the Association for Computational Linguistics: ACL (Findings of ACL), pp. 37278–37298. doi:10.18653/v1/2026.findings-acl.1857.

IEEE

Z. Yao et al., “Adaptive Backtracking for Privacy Protection in Large Language Models,” Findings of the Association for Computational Linguistics: ACL (Findings of ACL), pp. 37278–37298, 2026, doi: 10.18653/v1/2026.findings-acl.1857.

BibTeX

@inproceedings{yao2026adaptive,
  title     = {{Adaptive Backtracking for Privacy Protection in Large Language Models}},
  author    = {Yao, Zhihao and Gu, Yixin and Feng, Xiachong and Ma, Wenshuo and Li, Bo and Feng, Xiaocheng and Qin, Bing},
  booktitle = {Findings of the Association for Computational Linguistics: ACL (Findings of ACL)},
  year      = {2026},
  pages     = {37278--37298},
  doi       = {10.18653/v1/2026.findings-acl.1857},
  url       = {https://doi.org/10.18653/v1/2026.findings-acl.1857}
}

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

  • privacy leakage
  • large language models
  • retrieval-augmented generation
  • backtracking
  • personally identifiable information
  • adversarial attacks

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Yao, Z., Gu, Y., Feng, X., Ma, W., Li, B., Feng, X., & Qin, B. (2026). Adaptive Backtracking for Privacy Protection in Large Language Models. Findings of the Association for Computational Linguistics: ACL (Findings of ACL), 37278–37298. https://doi.org/10.18653/v1/2026.findings-acl.1857

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?

Zhihao Yao, Yixin Gu, Xiachong Feng, Wenshuo Ma, Bo Li, Xiaocheng Feng, Bing Qin.

What is the DOI of this paper?

10.18653/v1/2026.findings-acl.1857, permanently at https://doi.org/10.18653/v1/2026.findings-acl.1857. 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.