Preprint · 2026 · Open access

MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents

Cheng Zhang, Xuanning Cao, Zhicheng Zheng, Hui Xiong

arXiv (Cornell University) · Published 15 October 2026 · DOI 10.48550/arxiv.2610.21405

Open PDFView at the publisherCite this paper

In brief

“MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents” is a 2026 preprint by Cheng Zhang and 3 co-authors, published in arXiv (Cornell University). OpenAlex counts 1 citations as of 19 November 2026. It is open access: the full text is free to read at arXiv (Cornell University).

Citations
1
Most cited in
2026 (1)
References
25
Authors
4

Abstract

Long-term memory has become a crucial component of large language model (LLM) agents, with most research focusing on retrieving relevant information accurately and efficiently. However, a dependable memory system must also forget sensitive user data upon request. To evaluate this capability, we introduce MemCut, a systematic benchmark for user-requested forgetting. Its 456 hand-designed secrets in long-horizon conversations have fully annotated evidence trails to test recovery after deletion, including from messages that never state the secrets directly. We propose two recovery metrics. Native recovery measures reader success through the system's retrieval, whereas residual recovery measures the same reader's success when directly given surviving objects traced to annotated disclosures or evidence. We establish three reference deletion baselines to test how far existing retrievers and write models support forgetting across five memory designs and four LLMs. Disclosure removal reduces native recovery far more than residual recovery. LLM-guided deletion of ten additional turns from twenty retrieval candidates leaves 39-76% residual recovery. Subset sweeps (144 facts, three systems, two models) show that widening the pool at a fixed budget reduces residual recovery more than deleting all original candidates. These sweeps identify candidate coverage as a bottleneck in the reference baselines.

Cite this paper

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

APA 7

Zhang, C., Cao, X., Zheng, Z., & Xiong, H. (2026). MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.21405

MLA 9

Zhang, Cheng, et al. “MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents.” arXiv (Cornell University), 2026. https://doi.org/10.48550/arxiv.2610.21405.

Chicago

Zhang, Cheng, Xuanning Cao, Zhicheng Zheng, and Hui Xiong. 2026. “MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.21405.

Harvard

Zhang, C. et al. (2026) ‘MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents’, arXiv (Cornell University). doi:10.48550/arxiv.2610.21405.

IEEE

C. Zhang, X. Cao, Z. Zheng, and H. Xiong, “MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents,” arXiv (Cornell University), 2026, doi: 10.48550/arxiv.2610.21405.

BibTeX

@misc{zhang2026memcut,
  title     = {{MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents}},
  author    = {Zhang, Cheng and Cao, Xuanning and Zheng, Zhicheng and Xiong, Hui},
  howpublished = {arXiv (Cornell University)},
  year      = {2026},
  doi       = {10.48550/arxiv.2610.21405},
  url       = {https://doi.org/10.48550/arxiv.2610.21405}
}

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

  • long-term memory
  • LLM agents
  • secret recovery
  • memory design
  • privacy

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?

Zhang, C., Cao, X., Zheng, Z., & Xiong, H. (2026). MemCut: Every Secret Leaves a Trail in the Long-Term Memory of LLM Agents. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.21405

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 arXiv (Cornell University): sign in and open the PDF from this page.

Who wrote this paper?

Cheng Zhang, Xuanning Cao, Zhicheng Zheng, Hui Xiong.

What is the DOI of this paper?

10.48550/arxiv.2610.21405, permanently at https://doi.org/10.48550/arxiv.2610.21405. 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.