# Where Identity Lives: Localized, Retain-Free Identity Unlearning in Multimodal Large Language Models

arXiv: https://arxiv.org/abs/2608.30649 · HTML: https://mlsanity.com/paper/2608.30649/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 17 of its 28 references that are in the map.
- First posted on arXiv 31 Aug 2026.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM jailbreaks and adversarial robustness](https://mlsanity.com/mllm/mllm-jailbreaks-and-adversarial-robustness/)
- LLM safety and security / [Machine unlearning](https://mlsanity.com/safety/machine-unlearning/)

## Builds on (most cited first, 17 of 17)

- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [Machine Unlearning](https://mlsanity.com/paper/1912.03817/) (2019)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearning](https://mlsanity.com/paper/2404.05868/) (2024)
- [Unrolling SGD: Understanding Factors Influencing Machine Unlearning](https://mlsanity.com/paper/2109.13398/) (2021)
- [Does Localization Inform Editing? Surprising Differences in Causality-Based Localization vs. Knowledge Editing in Language Models](https://mlsanity.com/paper/2301.04213/) (2023)
- [Protecting Privacy in Multimodal Large Language Models with MLLMU-Bench](https://mlsanity.com/paper/2410.22108/) (2024)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/2507.10485/) (2025)
- [CLEAR: Character Unlearning in Textual and Visual Modalities](https://mlsanity.com/paper/2410.18057/) (2024)
- [The European Union general data protection regulation: what it is and what it means](https://mlsanity.com/paper/2510.02861/) (2025)
- [Causal Tracing of Object Representations in Large Vision Language Models: Mechanistic Interpretability and Hallucination Mitigation](https://mlsanity.com/paper/2511.05923/) (2025)
- [Dynamic Weight Grafting: Localizing Finetuned Factual Knowledge in Transformers](https://mlsanity.com/paper/2506.20746/) (2025)
- [Visual-Guided Key-Token Regularization for Multimodal Large Language Model Unlearning](https://mlsanity.com/paper/2601.22020/) (2026)
- [Before Forgetting, Learn to Remember: Revisiting Foundational Learning Failures in LVLM Unlearning Benchmarks](https://mlsanity.com/paper/2605.03759/) (2026)
- [AIM: Anchor Identity Features, Then Match for Multimodal Large Language Model Unlearning](https://mlsanity.com/paper/2608.28312/) (2026)

## Built on by (most cited first, 1 of 1)

- [SIEVE: Selective attention-value Suppression for Vision-Language Models Unlearning](https://mlsanity.com/paper/2610.01962/) (2026)

## Abstract (arXiv)

Removing a specific individual's information from multimodal large language models (MLLMs) is often needed after deployment, but existing methods rely on a retain set, which is hardest to obtain at that point, and rebuilding it recreates the privacy exposure that unlearning aims to remove. Forgetting from the forget set alone instead damages the shared visual-language computation, harming perception. We cast retain-free unlearning as a localization problem: causal tracing, weight transplant, and Fisher overlap all point to early-to-mid decoder MLPs as the layers where identity information is stored and, unlike other module families, can be modified without substantially disrupting vision. We turn this into Pathway-Aware Visual-attribute Anchoring (PAVA), which confines updates to these layers and pairs a forget loss with a visual-attribute anchor that preserves image-grounded behavior by distilling the model's own pre-unlearning answers from the forget images alone. On MLLMU-Bench and ReMem, PAVA gives the strongest forget-retain trade-off among forget-set-only methods and remains competitive with retain-based baselines.