LLM Unlearning

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

20 papers in the last four weeks, up 150% on the four weeks before. 0.2% of all new papers.

Jul 13Week of Sep 28

Latest papers 129

All topics

Backdoor Decontamination Dynamics in LLM Agents
Aug 11, 2026Gabriel Huang, Abhay Puri, Léo Boisvert +4LLM Backdoor Attacks Backdoor Defense in LLMs
ServiceNow Research · Mila – Quebec AI Institute · Polytechnique Montréal +3
OBLIVION: Workflow-Level Operational Skill Unlearning for Deployed Agents
Aug 8, 2026Zhengyang Shan, Xu Qian, Jiayun Xin +3AI Agent Security Benchmarks LLM Agent Safety
Shandong University
GROM: Gradient-Free Rapid One-Shot Machine Unlearning
Aug 6, 2026Paweł Batorski, Przemysław Spurek, Paul SwobodaMachine Unlearning LLM Unlearning
1Heinrich Heine University Düsseldorf · 2Jagiellonian University · 3IDEAS Research Institute
A Model Merging Approach for Continual MLLM Unlearning
Aug 5, 2026Yuhang Wang, Linlin Zhang, Haoxuan Ji +3Multimodal Large Language Models MLLM Unlearning
1Xidian University, Xi’an, China · 2Xi’an Jiaotong University, Xi’an, China
Leak-Resistant Unlearning: A New Benchmark for Evaluating Multi-Hop Reasoning Consistency and Recovery Robustness
Aug 5, 2026Haoting Qian, Qingjie Zhang, Zhicong Huang +2Multi-Hop Reasoning Language Model Robustness
Tsinghua University, China. · Ant Group, China.
Trajectory-Guided Forget-Recover Network for Continual LLM Unlearning
Aug 4, 2026Zezheng Wu, Xinghe Cheng, Qinggang Zhang +4Machine Unlearning LLM Unlearning
Guilin University of Electronic Technology · Jinan University · Jilin University +2
Exploring and Bridging Knowledge Holes in Unlearned Multimodal Large Language Models
Aug 3, 2026Junxiang You, Junkai Chen, Yuhao He +3Multimodal Large Language Models MLLM Unlearning
University of Chinese Academy of Sciences · Institute of Automation, Chinese Academy of Sciences · 3Cloudspace Technology
QR-Erase: Efficient Subspace-Based Machine Unlearning with Layer Localization
Aug 2, 2026Tyler Lizzo, Larry HeckMachine Unlearning LLM Unlearning
Georgia Institute of Technology
Beyond Binary Rewards: A Comparative Study of Reward Design for Reinforcement Unlearning
Jul 30, 2026Efstratios Zaradoukas, Davide Gabrielli, Bardh Prenkaj +1Reinforcement Learning with Verifiable Rewards Machine Unlearning
Technical University of Munich, Germany · Munich Center for Machine Learning, Germany · Sapienza University of Rome, Italy
Crossing the Margin Cliff: Toward Relearn-Robust LLM Unlearning via Margin Calibration
Jul 30, 2026Xiangyu Yin, Jiaxu Liu, Zhen Chen +1Privacy Leakage in Language Models LLM Unlearning
Chalmers University of Technology, Gothenburg, Sweden · Imperial College London, London, United Kingdom · University of Liverpool, Liverpool, United Kingdom +1
Subtract, Transport, or Replay? Auditable Deletion from Language-Model Memory
Jul 30, 2026Vishwajith RameshMemory-Augmented Language Models LLM Auditing
Vy Labs, Inc.
Stochastic Meta-Unlearning: Bridging Language Backbone and Multimodal Unlearning
Jul 21, 2026Zijie Liu, Jinhao Duan, Gaowen Liu +2VLM Unlearning Vision-Language Models
UNC at Chapel Hill · Cisco Research · Michigan State University
Forget Narrowly, Retain Broadly: Unlearning as an Asymmetric Generalization Problem
Jul 10, 2026Amit Peleg, Naman Deep Singh, Naama Pearl +2Factual Knowledge in Language Models Machine Unlearning
Tübingen AI Center, University of Tübingen
POPS: Recovering Unlearned Multi-Modality Knowledge in MLLMs with Prompt-Optimized Parameter Shaking
Jul 7, 2026Zhangheng LI, Jianing Zhu, Junyuan Hong +4Multimodal Large Language Models Privacy Leakage in Language Models
University of Texas at Austin · DEVCOM Army Research Laboratory
MPSelectTune: Prompt-type Selection for Fine-tuning improves Concept Unlearning in LLMs
Jul 4, 2026Shubhadip Nag, Srinjoy Das, Agniva Saha +5LLM Fine-Tuning Machine Unlearning
*IIT Kharagpur · †MPI-SWS · ‡HPE Labs
LACUNA: A Testbed for Evaluating Localization Precision for LLM Unlearning
Jul 2, 2026Matteo Boglioni, Thibault Rousset, Siva Reddy +2Privacy Leakage in Language Models Machine Unlearning
Mila – Quebec Artificial Intelligence Institute · McGill University · McGill-NLP/LACUNA
Auditing Forgetting in Limited Memory Language Models
Jul 1, 2026Arya Raeesi, Hanna RoedMemory-Augmented Language Models LLM Auditing
University of California, Berkeley, Berkeley, California, United States of America.
Revocable Learned State via Process Sidecars
Jun 29, 2026John SweeneyLLM Safety Alignment Task Arithmetic
Sideplane AI
MemLeak: Diagnosing Information Leaks in Multimodal Agent Memory
Jun 29, 2026Kuan Wang, Chao ZhangAgent Memory Multimodal Memory
Georgia Institute of Technology
CBD: API-Only LLM Black-Box Unlearning through Controlled Behavioral Divergence
Jun 26, 2026Zhiqiang Xie, Yijing Lin, Zhipeng Gao +1Machine Unlearning LLM Unlearning
State Key Laboratory of Networking and Switching Technology, Beijing University of Posts and Telecommunications · Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing · Department of Electronic and Electrical Engineering, Sungkyunkwan University, Suwon, South Korea
Natural Ungrokking: Asymmetric Control of Which Rules Survive Pretraining
Jun 24, 2026Juliana Li, Diya SreedharLanguage Model Pretraining LLM Unlearning
Harvard University, Cambridge, MA, USA.
Detect, Unlearn, Restore: Defending Text Summarization Models Against Data Poisoning
Jun 24, 2026Poojitha Thota, Shirin NilizadehLLM Auditing LLM Fine-Tuning
The University of Texas at Arlington Arlington, Texas, USA
Uncertainty-based Debiasing and Unlearning for Decontamination
Jun 22, 2026Guangzhi Sun, Xiao Zhan, Mark GalesData Contamination in Language Models LLM Unlearning
Department of Engineering, University of Cambridge · VRAIN, Universitat Politècnica de València
Mechanism-Guided Selective Unlearning for RLVR-Induced Reasoning
Jun 17, 2026Chenyu Zhou, Qiliang Jiang, Shuning Wu +1Machine Unlearning LLM Unlearning
School of Engineering, Institute of Science Tokyo, Japan · College of Control Science and Engineering, Zhejiang University, China · Department of Electrical and Computer Engineering, National University of Singapore, Singapore
Output Vector Editing for Memorization Mitigation in Large Language Models
Jun 17, 2026Ahmad Dawar Hakimi, Kaiwei Lei, Isabelle Augenstein +1Memorization in Language Models Neural Network Memorization
Center for Information and Language Processing, LMU Munich, Germany · Department of Computer Science, University of Copenhagen, Denmark · 3Munich Center for Machine Learning, Germany +1
PreUnlearn: Auditing Collateral Knowledge Damage Before Large Language Model Unlearning
Jun 16, 2026Bo Su, Ankit Shah, Thai LeLLM Auditing Machine Unlearning
Indiana University Bloomington, IN, USA
SAGE: Retain-Aware Post-Hoc Sanitization of Final Unlearning Vector
Jun 16, 2026Jingyuan Zhang, Yucheng Bai, Peixi Wen +6Machine Unlearning LLM Unlearning
Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University
RepSelect: Robust LLM Unlearning via Representation Selectivity
Jun 15, 2026Filip Sondej, Yushi Yang, Adam MahdiLLM Safety Representation Geometry in Language Models
Independent · University of Oxford
Auditing Machine Unlearning: A Systematic Research on Whether Models Truly Forget
Jun 15, 2026Dayong Ye, Tianqing Zhu, Ruiding Huang +5Machine Unlearning LLM Unlearning
Replay What Matters: Off-Policy Replay for Efficient LLM Reinforcement Unlearning
Jun 13, 2026Zirui Pang, Chenlong Zhang, Haosheng Tan +3Experience Replay Machine Unlearning
The Hong Kong University of Science and Technology (Guangzhou) · Institute of Automation, Chinese Academy of Sciences · University of Glasgow

Aug 11, 2026, cs.CRJ/K move · Enter open · S save

Backdoor Decontamination Dynamics in LLM Agents

Gabriel Huang, Abhay Puri, Léo Boisvert, Alexandre Drouin+3

ServiceNow Research · Mila – Quebec AI Institute · Polytechnique Montréal+3

LLM Backdoor Attacks Backdoor Defense in LLMs LLM Agent Security LLM Security Backdoor Defense LLM Unlearning

Open paper AskPDF arXiv

Open-weight LLM agents are vulnerable to backdoors installed during fine-tuning, which may be undetectable if the trigger conditions are never met during testing. Assuming defenders do not know the existing trigger, they cannot unlearn it directly. One decontamination strategy is to install a known backdoor (defensive poisoning) then to unlearn it, hoping that the original unknown backdoor is removed as a side effect. However, this procedure has uncertain outcomes: the original backdoor may persist or be erased or rerouted, among other possibilities. We introduce a framework for studying these dynamics in tool-calling agents, decoupling trigger, response, teacher, and fine-tuning method across systematic experiments on AgentDyn. Across 115 experiments, defensive poisoning alone erases around 56% of original backdoors; subsequent decontamination then drives almost all survivors to erasure, confirming that trigger recognition and malicious execution are behaviorally dissociable. Interestingly, our experiments find that malicious backdoors never persist when using different triggers of the same general type as the defensive backdoor when followed by decontamination via unlearning. Co-installing up to four backdoors increases resistance (around 36% erased), yet decontaminating a single known co-resident backdoor collaterally clears 52/60 co-residents (87%). Upon visualizing postdecontamination model internals using J-lens, we confirm that although the decontamination restores benign LLM responses, traces of original trigger awareness persist at intermediate layers.

Momentum

Related topics

Latest papers 129