mlsanityData to 7 Oct 2026

PPU-Bench:Real World Benchmark for Personalized Partial Unlearning in Vision Language Models

arXiv 2605.08800 ↗First posted 9 May 2026cs.CV cs.AIcited by 1

On the map: Multimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness

also in 1 more
Abstract (arXiv)

Multimodal Large Language Models (MLLMs) may memorize sensitive cross-modal information during pretraining. However, existing MLLM unlearning benchmarks rely on synthetic knowledge injection or complete subject-level deletion, which fail to capture realistic, personalized deletion requests that require fine-grained factual control. In this paper, we introduce PPU-Bench, a real-world and fine-tuning-free benchmark for personalized partial unlearning in MLLMs. PPU-Bench contains 24K multimodal and unimodal samples derived from pre-existing knowledge of 500 public figures under three progressively challenging settings: Complete, Selective, and Personalized unlearning. The benchmark evaluates whether methods can remove target knowledge while preserving non-target facts, model utility, and cross-modal consistency. Extensive experiments show that Complete Unlearning often suppresses visual identity rather than factual knowledge, while Selective and Personalized Unlearning expose significant forget--retain trade-offs and challenges in intra-subject factual boundaries. Robustness analysis under cross-image and prompt-based attacks reveals distinct vulnerabilities across different unlearning settings. Motivated by these findings, we propose Boundary-Aware Optimization (BAO), which explicitly models intra-subject forget-retain boundaries. Experimental results on two representative methods demonstrate that BAO can effectively enforce intra-subject factual boundaries.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 21 of its 43 references that are in the map. First posted on arXiv 9 May 2026.

Builds on

21 of 43 references.

Closest ancestors · 7

Same direction, newest first

  1. 2025OFFSIDE: Benchmarking Unlearning Misinformation in Multimodal Large Language Models26 Oct
  2. 2025PEBench: A Fictitious Dataset to Benchmark Machine Unlearning for Multimodal Large Language Models16 Mar
  3. 2025Modality-Aware Neuron Pruning for Unlearning in Multimodal Large Language Models21 Feb
  4. 2025MMUnlearner: Reformulating Multimodal Machine Unlearning in the Era of Multimodal Large Language Models16 Feb
  5. 2024Benchmarking Vision Language Model Unlearning via Fictitious Facial Identity Dataset5 Nov
  6. 2024Protecting Privacy in Multimodal Large Language Models with MLLMU-Bench29 Oct
  7. 2024Single Image Unlearning: Efficient Machine Unlearning in Multimodal Large Language Models21 May

Other references · 14

  1. 2023MMBench: Is Your Multi-modal Model an All-around Player?cited by 2,105
  2. 2024TOFU: A Task of Fictitious Unlearning for LLMscited by 507
  3. 2024Negative Preference Optimization: From Catastrophic Collapse to Effective Unlearningcited by 484
  4. 2021Unrolling SGD: Understanding Factors Influencing Machine Unlearningcited by 274
  5. 2024Towards Safer Large Language Models through Machine Unlearningcited by 134
  6. 2022Continual Learning and Private Unlearningcited by 124
  7. 2024RWKU: Benchmarking Real-World Knowledge Unlearning for Large Language Modelscited by 77
  8. 2024Machine Unlearning in Generative AI: A Surveycited by 49
  9. 2024CLEAR: Character Unlearning in Textual and Visual Modalitiescited by 36
  10. 2025Evaluating and Steering Modality Preferences in Multimodal Large Language Modelcited by 33
  11. 2024Generated Data with Fake Privacy: Hidden Dangers of Fine-tuning Large Language Models on Generated Datacited by 21
  12. 2024Benchmarking and Improving Large Vision-Language Models for Fundamental Visual Graph Understanding and Reasoningcited by 17
2 more
  1. 2025MMPB: It's Time for Multi-Modal Personalizationcited by 14
  2. 2025Praxis-VLM: Vision-Grounded Decision Making via Text-Driven Reinforcement Learningcited by 9

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Open-MMUnlearning: Unifying Methods and Evaluation for MLLM Unlearningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Stochastic Meta-Unlearning: Bridging Language Backbone and Multimodal Unlearning
  2. 2026PRMU: A Corpus-Free Benchmark for Person-Centric Knowledge Unlearning in Multimodal Large Language Models
  3. 2026Visual-Noise Guided In-Context Distillation for Multimodal Large Language Model Unlearning
  4. 2026Knowledge Vector Weakening: Efficient Training-free Unlearning for Large Vision-Language Models
  5. 2026PPE-Bench: A Benchmark for Evaluating MLLM Unlearning under Private-Public Entanglement
  6. 2025MLLMEraser: Achieving Test-Time Unlearning in Multimodal Large Language Models through Activation Steering
  7. 2026On the Robustness of Machine Unlearning for Vision-Language Models
  8. 2026Multimodal Unlearning Across Vision, Language, Video, and Audio: Survey of Methods, Datasets, and Benchmarks