# Evaluating Deep Unlearning in Large Language Models

arXiv: https://arxiv.org/abs/2410.15153 · HTML: https://mlsanity.com/paper/2410.15153/

## Key facts

- Cited by 14 papers in the map (data through 14 Oct 2026).
- Placed in Machine unlearning, one of 12 directions in LLM safety and security.
- Builds on 45 of its 67 references that are in the map.
- First posted on arXiv 19 Oct 2024.

## Placed in

- LLM safety and security / [Machine unlearning](https://mlsanity.com/safety/machine-unlearning/)

## Builds on (most cited first, 30 of 45)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Direct Preference Optimization: Your Language Model is Secretly a Reward Model](https://mlsanity.com/paper/2305.18290/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [PIQA: Reasoning about Physical Commonsense in Natural Language](https://mlsanity.com/paper/1911.11641/) (2019)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [HuggingFace's Transformers: State-of-the-art Natural Language Processing](https://mlsanity.com/paper/1910.03771/) (2019)
- [Extracting Training Data from Large Language Models](https://mlsanity.com/paper/2012.07805/) (2020)
- [Editing Models with Task Arithmetic](https://mlsanity.com/paper/2212.04089/) (2022)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [Embedding Entities and Relations for Learning and Inference in Knowledge Bases](https://mlsanity.com/paper/1412.6575/) (2014)
- [RACE: Large-scale ReAding Comprehension Dataset From Examinations](https://mlsanity.com/paper/1704.04683/) (2017)
- [Knowledge Graphs](https://mlsanity.com/paper/2003.02320/) (2020)
- [Textbooks Are All You Need II: phi-1.5 technical report](https://mlsanity.com/paper/2309.05463/) (2023)
- [Editing Factual Knowledge in Language Models](https://mlsanity.com/paper/2104.08164/) (2021)
- [Certified Data Removal from Machine Learning Models](https://mlsanity.com/paper/1911.03030/) (2019)
- [The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning](https://mlsanity.com/paper/2403.03218/) (2024)
- [Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning](https://mlsanity.com/paper/2404.05868/) (2024)
- [Making AI Forget You: Data Deletion in Machine Learning](https://mlsanity.com/paper/1907.05012/) (2019)
- [Knowledge Unlearning for Mitigating Privacy Risks in Language Models](https://mlsanity.com/paper/2210.01504/) (2022)
- [Scalable Extraction of Training Data from (Production) Language Models](https://mlsanity.com/paper/2311.17035/) (2023)
- [Neural Probabilistic Logic Programming in DeepProbLog](https://mlsanity.com/paper/1907.08194/) (2019)
- [Who's Harry Potter? Approximate Unlearning in LLMs](https://mlsanity.com/paper/2310.02238/) (2023)
- [StructGPT: A General Framework for Large Language Model to Reason over Structured Data](https://mlsanity.com/paper/2305.09645/) (2023)
- [Large Language Model Unlearning](https://mlsanity.com/paper/2310.10683/) (2023)
- [MQuAKE: Assessing Knowledge Editing in Language Models via Multi-Hop Questions](https://mlsanity.com/paper/2305.14795/) (2023)
- [A Review of Relational Machine Learning for Knowledge Graphs](https://mlsanity.com/paper/1503.00759/) (2015)
- [MUSE: Machine Unlearning Six-Way Evaluation for Language Models](https://mlsanity.com/paper/2407.06460/) (2024)
- [Rethinking Machine Unlearning for Large Language Models](https://mlsanity.com/paper/2402.08787/) (2024)

## Built on by (most cited first, 14 of 14)

- [Do LLMs Really Forget? Evaluating Unlearning with Knowledge Correlation and Confidence Awareness](https://mlsanity.com/paper/2506.05735/) (2025)
- [ForgetMe: Evaluating Selective Forgetting in Generative Models](https://mlsanity.com/paper/2504.12574/) (2025)
- [Agents Are All You Need for LLM Unlearning](https://mlsanity.com/paper/2502.00406/) (2025)
- [Keeping an Eye on LLM Unlearning: The Hidden Risk and Remedy](https://mlsanity.com/paper/2506.00359/) (2025)
- [LUNE: Efficient LLM Unlearning via LoRA Fine-Tuning with Negative Examples](https://mlsanity.com/paper/2512.07375/) (2025)
- [The Trojan Knowledge: Bypassing Commercial LLM Guardrails via Harmless Prompt Weaving and Adaptive Tree Search](https://mlsanity.com/paper/2512.01353/) (2025)
- [On the Impossibility of Retrain Equivalence in Machine Unlearning](https://mlsanity.com/paper/2510.16629/) (2025)
- [Erasing Without Remembering: Implicit Knowledge Forgetting in Large Language Models](https://mlsanity.com/paper/2502.19982/) (2025)
- [ThinkEval: Practical Evaluation of Knowledge Leakage in LLM Editing using Thought-based Knowledge Graphs](https://mlsanity.com/paper/2506.01386/) (2025)
- [Learning-Time Encoding Shapes Unlearning in LLMs](https://mlsanity.com/paper/2506.15076/) (2025)
- [WaterDrum: Watermarking for Data-centric Unlearning Metric](https://mlsanity.com/paper/2505.05064/) (2025)
- [Behemoth: Benchmarking Unlearning in LLMs Using Fully Synthetic Data](https://mlsanity.com/paper/2601.23153/) (2026)
- [What to Forget in Unlearning? Forget Set Curation for Language Models](https://mlsanity.com/paper/2608.14855/) (2026)
- [Bounding Retraining Equivalence and the Deletion Floor in Materials Machine Unlearning](https://mlsanity.com/paper/2609.35635/) (2026)

## Abstract (arXiv)

Machine unlearning has emerged as an important component in developing safe and trustworthy models. Prior work on fact unlearning in LLMs has mostly focused on removing a specified target fact robustly, but often overlooks its deductive connections to other knowledge. We propose a new setting for fact unlearning, deep unlearning, where the goal is not only to remove a target fact but also to prevent it from being deduced via retained knowledge in the LLM and logical reasoning. We propose three novel metrics: Success-DU and Recall to measure unlearning efficacy, and Accuracy to measure the remainder model utility. To benchmark this setting, we leverage both (1) an existing real-world knowledge dataset, MQuAKE, that provides one-step deduction instances, and (2) newly construct a novel semi-synthetic dataset, Eval-DU, that allows multiple steps of realistic deductions among synthetic facts. Experiments reveal that current methods struggle with deep unlearning: they either fail to deeply unlearn, or excessively remove unrelated facts. Our results suggest that targeted algorithms may have to be developed for robust/deep fact unlearning in LLMs.