$\textbf{AGT$^{AO}$}$: Robust and Stabilized LLM Unlearning via Adversarial Gating Training with Adaptive Orthogonality

• URL: http://arxiv.org/abs/2602.01703v1
• Date: Mon, 02 Feb 2026 06:19:27 GMT
• Title: $\textbf{AGT$^{AO}$}$: Robust and Stabilized LLM Unlearning via Adversarial Gating Training with Adaptive Orthogonality
• Authors: Pengyu Li, Lingling Zhang, Zhitao Gao, Yanrui Wu, Yuxuan Dong, Huan Liu, Bifan Wei, Jun Liu,
• Abstract summary: We propose a unified framework designed to reconcile robust erasure with utility preservation.<n>Adversarial Gating Training (AGT)$ formulates unlearning as a latent-space min-max game.<n>Experiments demonstrate that AGT$ mitigates a superior trade-off between unlearning efficacy and model utility.
• Score: 28.07698632768221
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: While Large Language Models (LLMs) have achieved remarkable capabilities, they unintentionally memorize sensitive data, posing critical privacy and security risks. Machine unlearning is pivotal for mitigating these risks, yet existing paradigms face a fundamental dilemma: aggressive unlearning often induces catastrophic forgetting that degrades model utility, whereas conservative strategies risk superficial forgetting, leaving models vulnerable to adversarial recovery. To address this trade-off, we propose $\textbf{AGT$^{AO}$}$ (Adversarial Gating Training with Adaptive Orthogonality), a unified framework designed to reconcile robust erasure with utility preservation. Specifically, our approach introduces $\textbf{Adaptive Orthogonality (AO)}$ to dynamically mitigate geometric gradient conflicts between forgetting and retention objectives, thereby minimizing unintended knowledge degradation. Concurrently, $\textbf{Adversarial Gating Training (AGT)}$ formulates unlearning as a latent-space min-max game, employing a curriculum-based gating mechanism to simulate and counter internal recovery attempts. Extensive experiments demonstrate that $\textbf{AGT$^{AO}$}$ achieves a superior trade-off between unlearning efficacy (KUR $\approx$ 0.01) and model utility (MMLU 58.30). Code is available at https://github.com/TiezMind/AGT-unlearning.

Related papers

• OFMU: Optimization-Driven Framework for Machine Unlearning [5.100622189286672]
  Large language models increasingly require the ability to unlearn specific knowledge, such as user requests, copyrighted materials, or outdated information.<n>We propose OFMU, a penalty-based bi-level optimization framework that explicitly prioritizes forgetting while preserving retention.<n>We show that OFMU consistently outperforms existing unlearning methods in both efficacy and retained utility.
  arXiv  Detail & Related papers  (2025-09-26T15:31:32Z)
• Steerable Adversarial Scenario Generation through Test-Time Preference Alignment [58.37104890690234]
  Adversarial scenario generation is a cost-effective approach for safety assessment of autonomous driving systems.<n>We introduce a new framework named textbfSteerable textbfAdversarial scenario textbfGEnerator (SAGE)<n>SAGE enables fine-grained test-time control over the trade-off between adversariality and realism without any retraining.
  arXiv  Detail & Related papers  (2025-09-24T13:27:35Z)
• ESLM: Risk-Averse Selective Language Modeling for Efficient Pretraining [53.893792844055106]
  Large language model pretraining is compute-intensive, yet many tokens contribute marginally to learning, resulting in inefficiency.<n>We introduce Selective Efficient Language Modeling, a risk-aware algorithm that improves training efficiency and distributional robustness by performing online token-level batch selection.<n> Experiments on GPT-2 pretraining show that ESLM significantly reduces training FLOPs while maintaining or improving both perplexity and downstream performance compared to baselines.
  arXiv  Detail & Related papers  (2025-05-26T12:23:26Z)
• UniErase: Towards Balanced and Precise Unlearning in Language Models [69.04923022755547]
  Large language models (LLMs) require iterative updates to address the outdated information problem.<n>UniErase is a novel unlearning framework that demonstrates precision and balanced performances between knowledge unlearning and ability retaining.
  arXiv  Detail & Related papers  (2025-05-21T15:53:28Z)
• FLARE: Robot Learning with Implicit World Modeling [87.81846091038676]
  $textbfFLARE$ integrates predictive latent world modeling into robot policy learning.<n>$textbfFLARE$ achieves state-of-the-art performance, outperforming prior policy learning baselines by up to 26%.<n>Our results establish $textbfFLARE$ as a general and scalable approach for combining implicit world modeling with high-frequency robotic control.
  arXiv  Detail & Related papers  (2025-05-21T15:33:27Z)
• No Query, No Access [50.18709429731724]
  We introduce the textbfVictim Data-based Adrial Attack (VDBA), which operates using only victim texts.<n>To prevent access to the victim model, we create a shadow dataset with publicly available pre-trained models and clustering methods.<n>Experiments on the Emotion and SST5 datasets show that VDBA outperforms state-of-the-art methods, achieving an ASR improvement of 52.08%.
  arXiv  Detail & Related papers  (2025-05-12T06:19:59Z)
• FALCON: Fine-grained Activation Manipulation by Contrastive Orthogonal Unalignment for Large Language Model [34.2978692666639]
  We propose Fine-grained Activation manipuLation by Contrastive Orthogonal uNalignment (FALCON) as a representation-guided unlearning approach.<n>FALCON achieves superior unlearning effectiveness while maintaining model utility, exhibiting robust resistance against knowledge recovery attempts.
  arXiv  Detail & Related papers  (2025-02-03T16:05:15Z)
• $\ abla τ$: Gradient-based and Task-Agnostic machine Unlearning [7.04736023670375]
  We introduce Gradient-based and Task-Agnostic machine Unlearning ($nabla tau$) $nabla tau$ applies adaptive gradient ascent to the data to be forgotten while using standard gradient descent for the remaining data. We evaluate our framework's effectiveness using a set of well-established Membership Inference Attack metrics.
  arXiv  Detail & Related papers  (2024-03-21T12:11:26Z)
• Towards Robust Model-Based Reinforcement Learning Against Adversarial Corruption [60.958746600254884]
  This study tackles the challenges of adversarial corruption in model-based reinforcement learning (RL) We introduce an algorithm called corruption-robust optimistic MLE (CR-OMLE), which leverages total-variation (TV)-based information ratios as uncertainty weights for MLE. We extend our weighting technique to the offline setting, and propose an algorithm named corruption-robust pessimistic MLE (CR-PMLE)
  arXiv  Detail & Related papers  (2024-02-14T07:27:30Z)
• Robust Empirical Risk Minimization with Tolerance [24.434720137937756]
  We study the fundamental paradigm of (robust) $textitempirical risk minimization$ (RERM) We show that a natural tolerant variant of RERM is indeed sufficient for $gamma$-tolerant robust learning VC classes over $mathbbRd$.
  arXiv  Detail & Related papers  (2022-10-02T21:26:15Z)
• RelaxLoss: Defending Membership Inference Attacks without Losing Utility [68.48117818874155]
  We propose a novel training framework based on a relaxed loss with a more achievable learning target. RelaxLoss is applicable to any classification model with added benefits of easy implementation and negligible overhead. Our approach consistently outperforms state-of-the-art defense mechanisms in terms of resilience against MIAs.
  arXiv  Detail & Related papers  (2022-07-12T19:34:47Z)
• Hedging using reinforcement learning: Contextual $k$-Armed Bandit versus $Q$-learning [0.22940141855172028]
  We study the construction of replication strategies for contingent claims in the presence of risk and market friction. In this article, the hedging problem is viewed as an instance of a risk-averse contextual $k$-armed bandit problem. We find that the $k$-armed bandit model naturally fits to the Profit and Loss formulation of hedging.
  arXiv  Detail & Related papers  (2020-07-03T11:34:10Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.