# Agentic Unlearning: When LLM Agent Meets Machine Unlearning

Source: [Agentic Unlearning: When LLM Agent Meets Machine Unlearning](<https://www.sota2.com/research/paper/699bdf1a09c33900fcd804f1>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-04T05:00:05.078Z

Authors: Bin Wang, Fan Wang, Pingping Wang, Jinyu Cong, Yang Yu, Yilong Yin, Zhongyi Han, Benzheng Wei

Published: 2026-02-06T07:13:13.000Z

## Abstract

In this paper, we introduce \\textbf{agentic unlearning} which removes specified information from both model parameters and persistent memory in agents with closed-loop interaction. Existing unlearning methods target parameters alone, leaving two critical gaps: (i) parameter-memory backflow, where retrieval reactivates parametric remnants or memory artifacts reintroduce sensitive content, and (ii) the absence of a unified strategy that covers both parameter and memory pathways. We present Synchronized Backflow Unlearning (SBU), a framework that unlearns jointly across parameter and memory pathways. The memory pathway performs dependency closure-based unlearning that prunes isolated entities while logically invalidating shared artifacts. The parameter pathway employs stochastic reference alignment to guide model outputs toward a high-entropy prior. These pathways are integrated via a synchronized dual-update protocol, forming a closed-loop mechanism where memory unlearning and parametric suppression reinforce each other to prevent cross-pathway recontamination. Experiments on medical QA benchmarks show that SBU reduces traces of targeted private information across both pathways with limited degradation on retained data.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Machine Unlearning on MedQA QF=100](<https://www.sota2.com/research/sota/machine-unlearning-on-medqa-qf-100.md>): Machine Unlearning; MedQA QF=100; Forget Accuracy: 73.

- [Machine Unlearning on MedMCQA QF=1000](<https://www.sota2.com/research/sota/machine-unlearning-on-medmcqa-qf-1000.md>): Machine Unlearning; MedMCQA QF=1000; Forget Accuracy: 71.33.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2602.17692>)

[PDF](<https://arxiv.org/pdf/2602.17692.pdf>)