paperSeptember 2026Unreviewed

Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents

Chao Yao, Yangbo Wei, Zhen Huang, Junhong Qian, Chenle Chen, Shaoqiang Lu, Chen Wu, Lei He

Abstract

Long-running LLM agents are stateful: beyond the transcript they accrete compressed summaries, plaintext memory, pending tool plans, and, under every serving API, a KV cache. Yet today's "forget" operations delete a plaintext memory record and stop, leaving every artifact derived from the revoked information intact. We formalize execution-state unlearning: after a forget request, the agent must behave as if it had never observed the target. Modeling the runtime as a deterministic transition syst

Cite

@misc{yao2026forgetting,
  title = {{Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM Agents}},
  author = {Chao Yao and Yangbo Wei and Zhen Huang and Junhong Qian and Chenle Chen and Shaoqiang Lu and Chen Wu and Lei He},
  year = {2026},
  month = sep,
  eprint = {2609.04875},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2609.04875}
}