mlsanityData to 7 Oct 2026

Rethinking Machine Unlearning for Large Language Models

arXiv 2402.08787 ↗First posted 13 Feb 2024cs.LG cs.CLcited by 247

On the map: LLM reasoning → Reasoning for agents, planning and code

also in 1 more
Abstract (arXiv)

We explore machine unlearning (MU) in the domain of large language models (LLMs), referred to as LLM unlearning. This initiative aims to eliminate undesirable data influence (e.g., sensitive or illegal information) and the associated model capabilities, while maintaining the integrity of essential knowledge generation and not affecting causally unrelated information. We envision LLM unlearning becoming a pivotal element in the life-cycle management of LLMs, potentially standing as an essential foundation for developing generative AI that is not only safe, secure, and trustworthy, but also resource-efficient without the need of full retraining. We navigate the unlearning landscape in LLMs from conceptual formulation, methodologies, metrics, and applications. In particular, we highlight the often-overlooked aspects of existing LLM unlearning research, e.g., unlearning scope, data-model interaction, and multifaceted efficacy assessment. We also draw connections between LLM unlearning and related areas such as model editing, influence functions, model explanation, adversarial training, and reinforcement learning. Furthermore, we outline an effective assessment framework for LLM unlearning and explore its applications in copyright and privacy safeguards and sociotechnical harm reduction.

Key facts. Cited by 247 papers in the map (data through 7 Oct 2026). Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning. Builds on 126 of its 176 references that are in the map. First posted on arXiv 13 Feb 2024.

Builds on

126 of 176 references.

References · 126

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2022Training language models to follow instructions with human feedbackcited by 15,712
  3. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  4. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  5. 2017Deep reinforcement learning from human preferencescited by 3,846
  6. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 2,950
  7. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 2,914
  8. 2022Locating and Editing Factual Associations in GPTcited by 2,609
  9. 2022Emergent Abilities of Large Language Modelscited by 2,483
  10. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,221
  11. 2023Retrieval-Augmented Generation for Large Language Models: A Surveycited by 2,185
  12. 2020Extracting Training Data from Large Language Modelscited by 2,144
38 more
  1. 2017Understanding Black-box Predictions via Influence Functionscited by 1,714
  2. 2023Jailbroken: How Does LLM Safety Training Fail?cited by 1,565
  3. 2020Transformer Feed-Forward Layers Are Key-Value Memoriescited by 1,307
  4. 2018Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examplescited by 1,259
  5. 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,235
  6. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!cited by 1,106
  7. 2022Editing Models with Task Arithmeticcited by 1,076
  8. 2019Machine Unlearningcited by 1,020
  9. 2018The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networkscited by 847
  10. 2022Discovering Language Model Behaviors with Model-Written Evaluationscited by 805
  11. 2022Quantifying Memorization Across Neural Language Modelscited by 755
  12. 2023Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedbackcited by 743
  13. 2020Fast is better than free: Revisiting adversarial trainingcited by 723
  14. 2019Adversarial Training for Free!cited by 638
  15. 2023RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedbackcited by 615
  16. 2019Eternal Sunshine of the Spotless Net: Selective Forgetting in Deep Networkscited by 599
  17. 2021Knowledge Neurons in Pretrained Transformerscited by 590
  18. 2023Erasing Concepts from Diffusion Modelscited by 530
  19. 2023Are Emergent Abilities of Large Language Models a Mirage?cited by 515
  20. 2019Certified Data Removal from Machine Learning Modelscited by 512
  21. 2024TOFU: A Task of Fictitious Unlearning for LLMscited by 507
  22. 2023Jailbreaking ChatGPT via Prompt Engineering: An Empirical Studycited by 478
  23. 2024Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Trainingcited by 478
  24. 2023RRHF: Rank Responses to Align Language Models with Human Feedback without tearscited by 454
  25. 2019Making AI Forget You: Data Deletion in Machine Learningcited by 447
  26. 2022Memory-Based Model Editing at Scalecited by 443
  27. 2023Catastrophic Jailbreak of Open-source LLMs via Exploiting Generationcited by 439
  28. 2022Knowledge Unlearning for Mitigating Privacy Risks in Language Modelscited by 424
  29. 2023Scalable Extraction of Training Data from (Production) Language Modelscited by 415
  30. 2023SalUn: Empowering Machine Unlearning via Gradient-based Weight Saliency in Both Image Classification and Generationcited by 376
  31. 2023Detecting Pretraining Data from Large Language Modelscited by 375
  32. 2023Who's Harry Potter? Approximate Unlearning in LLMscited by 366
  33. 2021Remember What You Want to Forget: Algorithms for Machine Unlearningcited by 360
  34. 2023Fine-Tuning Language Models with Just Forward Passescited by 337
  35. 2023Ablating Concepts in Text-to-Image Diffusion Modelscited by 312
  36. 2023Studying Large Language Model Generalization with Influence Functionscited by 311
  37. 2020Descent-to-Delete: Gradient-Based Methods for Machine Unlearningcited by 303
  38. 2022A Survey of Machine Unlearningcited by 298

50 most cited of 126 All 126 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 60; 2025: 109; 2026: 78.

247 papers.

Sort:

By area:

  1. 2024Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearningcited by 484
  2. 2025Agentic Context Engineering: Evolving Contexts for Self-Improving Language Modelscited by 349
  3. 2024MUSE: Machine Unlearning Six-Way Evaluation for Language Modelscited by 251
  4. 2024Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMscited by 158
  5. 2024Tamper-Resistant Safeguards for Open-Weight LLMscited by 132
  6. 2024SAFREE: Training-Free and Adaptive Guard for Safe Text-to-Image And Video Generationcited by 125
  7. 2024A Practical Review of Mechanistic Interpretability for Transformer-Based Language Modelscited by 113
  8. 2024An Adversarial Perspective on Machine Unlearning for AI Safetycited by 113
  9. 2024Large Language Model Unlearning via Embedding-Corrupted Promptscited by 112
  10. 2024Guardrail Baselines for Unlearning in LLMscited by 90
  11. 2024Steering Language Model Refusal with Sparse Autoencoderscited by 89
  12. 2024Reversing the Forget-Retain Objectives: An Efficient LLM Unlearning Framework from Logit Differencecited by 85
38 more
  1. 2024Unlearning or Obfuscating? Jogging the Memory of Unlearned LLMs via Benign Relearningcited by 81
  2. 2025OpenUnlearning: Accelerating LLM Unlearning via Unified Benchmarking of Methods and Metricscited by 80
  3. 2024RWKU: Benchmarking Real-World Knowledge Unlearning for Large Language Modelscited by 77
  4. 2024Catastrophic Failure of LLM Unlearning via Quantizationcited by 76
  5. 2025Towards LLM Unlearning Resilient to Relearning Attacks: A Sharpness-Aware Minimization Perspective and Beyondcited by 76
  6. 2025STAIR: Improving Safety Alignment with Introspective Reasoningcited by 72
  7. 2024LLM Unlearning via Loss Adjustment with Only Forget Datacited by 71
  8. 2024Do Unlearning Methods Remove Information from Language Model Weights?cited by 66
  9. 2025Rethinking LLM Unlearning Objectives: A Gradient Perspective and Go Beyondcited by 59
  10. 2024Position: LLM Unlearning Benchmarks are Weak Measures of Progresscited by 58
  11. 2024What Makes and Breaks Safety Fine-tuning? A Mechanistic Studycited by 54
  12. 2024A Closer Look at Machine Unlearning for Large Language Modelscited by 54
  13. 2024On Large Language Model Continual Unlearningcited by 51
  14. 2024Challenging Forgets: Unveiling the Worst-Case Forget Sets in Machine Unlearningcited by 44
  15. 2024Single Image Unlearning: Efficient Machine Unlearning in Multimodal Large Language Modelscited by 44
  16. 2024Stress-Testing Capability Elicitation With Password-Locked Modelscited by 42
  17. 2025Exploring Criteria of Loss Reweighting to Enhance LLM Unlearningcited by 39
  18. 2025LLM Unlearning via Neural Activation Redirectioncited by 38
  19. 2024A Survey of Recent Backdoor Attacks and Defenses in Large Language Modelscited by 36
  20. 2025Model Tampering Attacks Enable More Rigorous Evaluations of LLM Capabilitiescited by 36
  21. 2024Threats, Attacks, and Defenses in Machine Unlearning: A Surveycited by 34
  22. 2024Offset Unlearning for Large Language Modelscited by 33
  23. 2024Advances and Open Challenges in Federated Foundation Modelscited by 33
  24. 2024Machine Unlearning: A Comprehensive Surveycited by 33
  25. 2024Unlearning as multi-task optimization: A normalized gradient difference approach with an adaptive learning ratecited by 29
  26. 2024Gradient Routing: Masking Gradients to Localize Computation in Neural Networkscited by 28
  27. 2024Competition Report: Finding Universal Jailbreak Backdoors in Aligned LLMscited by 27
  28. 2024Towards Effective Evaluations and Comparisons for LLM Unlearning Methodscited by 27
  29. 2024Adaptive Deployment of Untrusted LLMs Reduces Distributed Threatscited by 27
  30. 2025Security Concerns for Large Language Models: A Surveycited by 26
  31. 2024Towards Scalable Exact Machine Unlearning Using Parameter-Efficient Fine-Tuningcited by 25
  32. 2024WAGLE: Strategic Weight Attribution for Effective and Modular Unlearning in Large Language Modelscited by 25
  33. 2025Certified Unlearning for Neural Networkscited by 25
  34. 2024Unified Parameter-Efficient Unlearning for LLMscited by 24
  35. 2024Cross-Modal Safety Alignment: Is textual unlearning all you need?cited by 22
  36. 2024On Calibration of LLM-based Guard Models for Reliable Content Moderationcited by 21
  37. 2025SAEs Can Improve Unlearning: Dynamic Sparse Autoencoder Guardrails for Precision Unlearning in LLMscited by 21
  38. 2025Language Models May Verbatim Complete Text They Were Not Explicitly Trained Oncited by 20

50 most cited of 247 All 247 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025Model Collapse Is Not a Bug but a Feature in Machine Unlearning for LLMs
  2. 2026CAP: Controllable Alignment Prompting for Unlearning in LLMs
  3. 2024Towards Robust and Parameter-Efficient Knowledge Unlearning for LLMs
  4. 2024When Machine Unlearning Meets Retrieval-Augmented Generation (RAG): Keep Secret or Forget Knowledge?
  5. 2026Open-MMUnlearning: Unifying Methods and Evaluation for MLLM Unlearning
  6. 2024Machine Unlearning in Generative AI: A Survey
  7. 2026Trajectory Unlearning on LLM-based Agents
  8. 2026Explainable LLM Unlearning Through Reasoning