Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Towards Benign Memory Forgetting for Selective Multimodal Large Language Model Unlearning

T0 review · 4 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Selective MLLM unlearning can erase private image-text memories without sacrificing general image understanding, and the paper's SMFA method plus S-MLLMUn Bench demonstrate how to measure and achieve that balance.

desk verdict SMFA and S-MLLMUn Bench are a real step forward for MLLM unlearning, but the paper overclaims 'benign forgetting' — its own forget-set Fact Scores sit near 5/10, so erasure is partial, not precise. read the letter →

arxiv 2511.20196 v2 pith:2FPFEYS4 submitted 2025-11-25 cs.AI

Zhen Zeng , Leijiang Gu , Zhangling Duan , Feng Li , Cees G. M. Snoek , Meng Wang , Zenglin Shi This is my paper
classification cs.AI
keywords machineunlearningmultimodallargelanguagemodelsselectiveforgettingimageunderstandingparametermaskingrighttobeforgottenrefusaltuningbenchmark
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper targets a known failure of multimodal unlearning: methods that scrub private facts from an MLLM often damage its general ability to look at images and answer basic questions. It argues that the damage comes from over-generalization of the forgetting update, and that this can be filtered in weight space. It proposes SMFA, which fine-tunes the model on refusal labels to create a memory-forgetting adapter, then masks out adapter entries that conflict with a retaining anchor trained on a small set of knowledge to preserve. To make the trade-off measurable, it introduces S-MLLMUn Bench, a benchmark with 1,000 synthetic profiles pairing private image-memory questions with image-understanding questions. On two MLLMs, the paper reports that existing methods trade away image understanding, while SMFA forgets targeted knowledge and keeps understanding close to the original model.

What carries the argument

The load-bearing object is the Sculpted Memory Forgetting Adapter (SMFA), built from two weight updates. The Memory Forgetting Adapter ΔW_f comes from refusal-label fine-tuning on the forget set and encodes the forgetting effect. The Retaining Anchor ΔW_a comes from fine-tuning on a few-shot subset of the retain set and encodes the direction that preserves unrelated knowledge. The masking rule kills an entry of ΔW_f when the anchor and forgetting updates have opposite signs (directional conflict) and the forgetting entry exceeds |ρ·ΔW_a| (relative magnitude), with ρ = ||ΔW_f||_F / ||ΔW_a||_F making the two updates comparable. Merging the masked adapter back into the base model leaves the non

What would settle it

Train SMFA with a retain anchor built only from character-image questions, then measure ophthalmic image understanding on the forget set. If the few-shot anchor fails to generalize to the ophthalmic modality, that score should drop visibly after unlearning, showing the anchor does not protect unrelated capabilities. Alternatively, resample the few-shot retain set several times and check whether retain-set image understanding varies sharply across runs; high variance would falsify the anchor's reliability.

Watch

Extended reading notes

Core claim

The central claim is that benign forgetting—removing the targeted privacy-sensitive multimodal memory while preserving unrelated knowledge and foundational image understanding—is achievable by sculpting the forgetting update rather than by stronger forgetting. SMFA first obtains a Memory Forgetting Adapter ΔW_f by fine-tuning on forget-set questions with refusal labels, then builds a Retaining Anchor ΔW_a from a few-shot retain set. The mask removes entries of ΔW_f only when the two updates point in opposite directions and the forget entry is large relative to the anchor entry after scale normalization. The surviving update is merged into the base model. On S-MLLMUn Bench, SMFA drives Fact S

Load-bearing premise

The method depends on the claim that a few-shot retain set, matched in size to the forget set, produces a retaining anchor whose weight direction represents all unrelated knowledge and general visual understanding well enough that masking every conflicting dominant entry cannot delete anything important.

Editorial extensions

If this is right

  • If SMFA is right, MLLMs can be granted a practical right to be forgotten: targeted privacy erasure without retraining and without losing the vision-language skills that make the model useful.
  • The few-shot retain set (matched in size to the forget set, 5–15% of the data) means unlearning can proceed even when the full training corpus is unavailable.
  • S-MLLMUn Bench's joint evaluation of forgetting, retention, and image understanding gives the field a shared protocol for judging whether unlearning is benign rather than merely destructive.
  • The k hyperparameter gives practitioners a dial: raising k strengthens forgetting on the forget set while retain-set scores stay largely stable until k becomes very large.
  • The sign-and-magnitude masking rule connects unlearning to model merging, suggesting that conflicting updates can be reconciled in weight space without gradient ascent or neuron pruning.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper tests SMFA on synthetic profiles; a natural extension is to apply the same anchor-mask rule to naturally occurring private images such as faces, documents, and medical photos, where the anchor would need to generalize across less contrived visual variation.
  • A testable consequence of the anchor's 'strong generalization' assumption is that deliberately narrowing the few-shot retain set—for example, omitting an entire visual modality like the ophthalmic images—should make that modality vulnerable to unlearning; the paper's design could measure this directly.
  • The same weight-sculpting principle could plausibly apply to knowledge editing in MLLMs, where a factual correction is an update that should be confined rather than broadcast; the paper does not claim this, but the mechanism is structurally identical.
  • If benign forgetting becomes a standard objective, safety post-training could shift from broad refusal behavior toward identity-specific forgetting—models could describe a face generically while refusing to name or link it to private attributes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper introduces S-MLLMUn Bench, a benchmark for multimodal LLM unlearning that jointly evaluates forgetting of privacy-sensitive image-text memories and preservation of general image understanding. It uses 1,000 synthetic profiles with character and ophthalmic images. The authors propose SMFA: first fine-tune the model with refusal labels on the forget set to obtain a Memory Forgetting Adapter (MFA); then train a retaining anchor on a few-shot retain set and mask entries of the forget update that directionally conflict with the anchor and dominate it in magnitude (Eqs. 6–10). Experiments on LLaVA-OneVision-7B and Qwen2.5-VL-7B, against four baselines, show SMFA preserves image understanding and retain-set memory much better than baselines, while reducing forget-set Fact Scores from ~7.5 to ~4.7–5.3. The paper claims this is 'precise and controllable unlearning' and 'benign forgetting.'

Significance. The benchmark is a timely and useful attempt to quantify the forgetting-retention trade-off for MLLMs, and the sculpting idea is simple and plausible. Providing code/data is a plus. However, the forget-set Fact Scores are far above the base-model floor, so the central claim of targeted knowledge erasure is not supported by the paper's own measurements. The benchmark may still be valuable, but the method as presented does not yet substantiate 'benign forgetting' as rigorously defined.

major comments (4)
  1. [Table 1, all SMFA rows] Forget-set Fact Scores are 4.7–5.3 for image memory across models/ratios, compared to the base model's ~1.2–1.5 and the original's ~7.5. A mean of ~5 implies many responses still contain correct facts, not clean refusals. Thus the abstract's 'targeted knowledge erasure' and Sec. 5.2's 'effectively erases targeted knowledge' are not supported. Report the score distribution (e.g., percentage of Fact Score 0) and use the base model as the reference floor.
  2. [Sec. 5.1, Table 1] No error bars, multiple seeds, or significance tests are reported. The forget-set ratios are small and the few-shot retain set is equal in size, so the trade-off results could be unstable. Run at least three seeds and report mean±std, or justify determinism, before claiming superiority over baselines.
  3. [Sec. 3.3, Eqs. (6)–(10)] The retaining anchor is trained on a few-shot retain subset equal to the forget set. The paper relies on the MLLM's 'strong generalization capability' to propagate this limited signal, but no analysis shows how the choice of few-shot items affects the mask. Add an ablation varying the few-shot retain set size/random seed to demonstrate robustness of the anchor.
  4. [Sec. 2, Eq. (7)] The masking rule is close to TIES-Merging (trim and sign-conflict resolution), which is cited but not compared. Without a TIES-style baseline applied to the MFA and retaining anchor, the specific contribution of the 'sculpting' is unclear. Add such a comparison.
minor comments (5)
  1. [Table 1] Header reads 'LLaV A-OneVision'; should be 'LLaVA-OneVision'. Also the table is dense; consider splitting by model.
  2. [Fig. 6] Define 'F Score' and 'M Score' in the caption (Fact Score and Meaningful Score).
  3. [Sec. 5.2] Use 'Meaningful Score' consistently; the metric is named 'Meaningful Score' in Sec. 4.3 but appears in lower-case in some places.
  4. [App. A.1] State explicitly that the evaluation retain set is the complete retain set; the main text only says 'retain set.'
  5. [Sec. 4.3] Provide the exact prompt used for the Qwen-Plus Fact Score and Meaningful Score in the appendix to ensure reproducibility.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: SMFA's masking equations are self-contained heuristics, and the benchmark/method co-development is an external-validity concern, not a derivation-level circularity.

full rationale

The derivation chain is self-contained. The MFA update ΔW_f is obtained by fine-tuning on refusal-labeled forget data (Eqs. 4-5); the retaining anchor ΔW_a is obtained independently by fine-tuning on a few-shot retain subset; the mask M = C⊙R (Eqs. 6-9) is an explicit heuristic on sign conflicts and relative magnitudes; and the final update (Eq. 10) is a component-wise product. No reported evaluation quantity (Fact Score, ROUGE-L, Meaningful Score) appears in the objective or in the mask construction, and no equation is equal by construction to a benchmark result. The only overlapping-author citation [33] is non-load-bearing related work; the methodological debts to IDK, MANU, MLLMU-Bench, and LoRA are external. The co-developed benchmark and Qwen-family judge raise external-validity and self-evaluation concerns, and Table 1's forget-set Fact Scores around 4.7-5.3 undercut the 'precise erasure' claim, but these are correctness/evaluation risks, not circular reductions. Hence the score is minimal.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

SMFA relies on a few-shot anchor and linear-weight masking to separate forgetting from retention. The central free parameter is k=5; the main axioms are the generalizability of the few-shot anchor, the sufficiency of linear-layer editing, the representativeness of the synthetic benchmark, and the reliability of Qwen-Plus as a judge.

free parameters (2)
  • masking threshold k = 5
    Appears in Eq. (7) as the relative-magnitude threshold; set to 5 without a validation-based selection protocol. The parameter analysis in Sec 5.4 shows changing k controls the forgetting/retention tradeoff, so the headline result depends on this hand-chosen value.
  • LoRA and fine-tuning hyperparameters
    Learning rate, LoRA rank/alpha, number of steps, and optimizer are not specified in the paper, so an independent reproduction cannot know the exact recipe. These choices affect the behavior of both the MFA and the retaining anchor.
assumptions (4)
  • domain assumption A few-shot retain-set anchor generalizes to all retained knowledge and visual understanding.
    Sec. 3.3: 'the strong generalization capability of MLLMs enables this limited signal to propagate effectively, allowing delta-W_a to serve as a reliable anchor.' This is the central bet of the masking mechanism.
  • domain assumption Unlearning can be confined by editing only linear-layer LoRA weights.
    Sec. 3.2 and Sec. 5.1: all fine-tuning is on linear layers with LoRA, and the resulting delta is treated as a 'Memory Forgetting Adapter'. No evidence is given that cross-attention, norms, or embeddings are unnecessary for the forgetting/retention tradeoff.
  • domain assumption The synthetic S-MLLMUn Bench profiles and Qwen-generated questions capture real privacy-unlearning scenarios.
    Appendices A and A.1: all data are fictitious, generated from StyleGAN faces and Qwen-VL-Plus text, plus DeepEyeNet ophthalmic images. External validity to real user data is assumed.
  • domain assumption Qwen-Plus is an accurate judge of factual erasure and meaningfulness.
    Appendix A.2: Fact Score and Meaningful Score are produced by Qwen-Plus; the paper assumes this judge reliably measures both erasure and coherence.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Towards Benign Memory Forgetting for Selective Multimodal Large Language Model Unlearning." pith.science (2026). https://pith.science/paper/2FPFEYS4

@misc{pith2026251120196,
  author       = {Pith},
  title        = {Pith review of: Towards Benign Memory Forgetting for Selective Multimodal Large Language Model Unlearning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2FPFEYS4}},
  note         = {Machine review of arXiv:2511.20196}
}
read the original abstract

Multimodal large language models (MLLMs) can inadvertently memorize privacy-sensitive information during training. While existing unlearning methods can remove such content, they often severely degrade the model's foundational capabilities, such as general image understanding. This critical shortfall motivates our investigation into benign memory forgetting, the precise removal of targeted, privacy-sensitive knowledge while rigorously preserving unrelated capabilities. To pioneer and evaluate progress toward this objective, we introduce S-MLLMUn Bench, the first benchmark designed to jointly and quantitatively assess an unlearning method's efficacy in knowledge erasure and the preservation of image understanding. Furthermore, we propose the Sculpted Memory Forgetting Adapter (SMFA), a new framework that enables benign memory forgetting. SMFA confines forgetting to designated memory regions, maintaining overall model performance. By initially fine-tuning the model to replace sensitive outputs with refusals, SMFA generates a memory forgetting adapter, followed by a retaining anchor-guided masking mechanism that safeguards unrelated knowledge. Extensive experiments on S-MLLMUn Bench demonstrate that existing methods fail to achieve benign forgetting, whereas our proposed SMFA serves as an effective baseline, successfully achieving targeted knowledge erasure without compromising the model's foundational visual capabilities. Code and data are available at https://github.com/zeng-zhen/S-MLLMUn.

Figures

Figures reproduced from arXiv: 2511.20196 by the authors.

Figure 1
Figure 1. (a) The goal of MLLM unlearning is to make the model selectively forget image knowledge memory, while preserving its general [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed Sculpted Memory Forgetting Adapter (SMFA). First, a Memory Forgetting Adapter (MFA) is derived [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Overall pipeline of S-MLLMUn Bench. It includes a fine-tuning dataset, an unlearning dataset, and an evaluation dataset. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Analysis of the hyperparameter k on LLaVA-OneVision with forget ratio 5% and 10%. Orig. denotes Original. much more effectively. This advantage stems from our pre￾cise sculpting, which filters over-generalization forgetting updates while retaining beneficial ones, ther…
Figure 5
Figure 5. Figure 5: Comparison of image understanding ability across different image types under various unlearning methods on LLaVA-OneVision [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: A Case study on S-MLLMUn Bench. The example [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 7
Figure 7. Figure 7: Example of S-MLLMUn Bench. loss on the forget set. The overall objective is defined as: LKL = −L (Df , θ) + 1 [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Some examples of refusal labels. troduced to finally determine the pruned neurons based on previous importance function. Given a pruning rate α and Sn, MANU defines a pruned neurons set: N = { n : Sn is among the top α% of all scores }. For each neu￾ron n ∈ N , MANU se…
Figure 9
Figure 9. Figure 9: Outputs of models on image memory questions after unlearning with different methods. [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: Outputs of models on image understanding and text memory questions after unlearning with different methods. [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Toward Fine-Grained Forgetting:Attribute Unlearning for Multimodal Large Language Models

    cs.AI 2026-08 reject novelty 6.0 of 10

    The paper defines attribute-level MLLM unlearning and proposes CLRP, but the method's headline forgetting gains on cloze are partly produced by test-time logit subtraction applied only to the forget and test sets.

Reference graph

Works this paper leans on

35 extracted references · 17 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Gpt-4 technical report.arXiv preprint arXiv:2303.08774,

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ah- mad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report.arXiv preprint arXiv:2303.08774,

    arXiv

  2. [2]

    Flamingo: a visual language model for few-shot learning.NeurIPS, 2022

    Jean-Baptiste Alayrac, Jeff Donahue, Pauline Luc, Antoine Miech, Iain Barr, Yana Hasson, Karel Lenc, Arthur Mensch, Katherine Millican, Malcolm Reynolds, et al. Flamingo: a visual language model for few-shot learning.NeurIPS, 2022. 1

    2022

  3. [3]

    Palm 2 technical report.arXiv preprint arXiv:2305.10403, 2023

    Rohan Anil, Andrew M Dai, Orhan Firat, Melvin John- son, Dmitry Lepikhin, Alexandre Passos, Siamak Shakeri, Emanuel Taropa, Paige Bailey, Zhifeng Chen, et al. Palm 2 technical report.arXiv preprint arXiv:2305.10403, 2023. 1

    arXiv 2023

  4. [4]

    Qwen-vl: A frontier large vision-language model with versatile abilities.arXiv preprint arXiv:2308.12966, 1(2):3,

    Jinze Bai, Shuai Bai, Shusheng Yang, Shijie Wang, Sinan Tan, Peng Wang, Junyang Lin, Chang Zhou, and Jingren Zhou. Qwen-vl: A frontier large vision-language model with versatile abilities.arXiv preprint arXiv:2308.12966, 1(2):3,

    arXiv

  5. [5]

    Qwen2.5-vl technical report.arXiv preprint arXiv:2502.13923, 2025

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhao- hai Li, Jianqiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Jun- yang Lin. Qwen2.5-vl technical repor...

    arXiv 2025

  6. [6]

    Towards making systems for- get with machine unlearning

    Yinzhi Cao and Junfeng Yang. Towards making systems for- get with machine unlearning. InProceedings of the 2015 IEEE Symposium on Security and Privacy (SP), pages 463–

    2015

  7. [7]

    Palm: Scaling language modeling with pathways.Journal of Machine Learning Research, 24(240): 1–113, 2023

    Aakanksha Chowdhery, Sharan Narang, Jacob Devlin, Maarten Bosma, Gaurav Mishra, Adam Roberts, Paul Barham, Hyung Won Chung, Charles Sutton, Sebastian Gehrmann, et al. Palm: Scaling language modeling with pathways.Journal of Machine Learning Research, 24(240): 1–113, 2023. 1

    2023

  8. [8]

    Clear: Character un- learning in textual and visual modalities.arXiv preprint arXiv:2410.18057, 2024

    Alexey Dontsov, Dmitrii Korzh, Alexey Zhavoronkin, Boris Mikheev, Denis Bobkov, Aibek Alanov, Oleg Y Rogov, Ivan Oseledets, and Elena Tutubalina. Clear: Character un- learning in textual and visual modalities.arXiv preprint arXiv:2410.18057, 2024. 2

    arXiv 2024

Show all 35 references
  1. [9]

    Avoiding copyright infringement via machine unlearning.arXiv preprint arXiv:2406.10952, 2024

    Guangyao Dou, Zheyuan Liu, Qing Lyu, Kaize Ding, and Eric Wong. Avoiding copyright infringement via machine unlearning.arXiv preprint arXiv:2406.10952, 2024. 2

    2024 arXiv

  2. [10]

    Deepopht: medical report generation for retinal images via deep models and visual explanation

    Jia-Hong Huang, C-H Huck Yang, Fangyu Liu, Meng Tian, Yi-Chieh Liu, Ting-Wei Wu, I-Hung Lin, Kang Wang, Hiro- masa Morikawa, Hernghua Chang, Jesper Tegner, and Mar- cel Worring. Deepopht: medical report generation for retinal images via deep models and visual explanation. InPr...

    2021

  3. [11]

    Editing models with task arithmetic.arXiv preprint arXiv:2212.04089, 2022

    Gabriel Ilharco, Marco Tulio Ribeiro, Mitchell Wortsman, Suchin Gururangan, Ludwig Schmidt, Hannaneh Hajishirzi, and Ali Farhadi. Editing models with task arithmetic.arXiv preprint arXiv:2212.04089, 2022. 3

    2022 arXiv

  4. [12]

    A style-based generator architecture for generative adversarial networks

    Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019. 5

    2019

  5. [13]

    Llava-onevision: Easy visual task transfer

    Bo Li, Yuanhan Zhang, Dong Guo, Renrui Zhang, Feng Li, Hao Zhang, Kaichen Zhang, Yanwei Li, Ziwei Liu, and Chunyuan Li. Llava-onevision: Easy visual task transfer. arXiv preprint arXiv:2408.03326, 2024. 6

    2024 arXiv

  6. [14]

    Sin- gle image unlearning: Efficient machine unlearning in multi- modal large language models.Advances in Neural Informa- tion Processing Systems, 37:35414–35453, 2024

    Jiaqi Li, Qianshan Wei, Chuanyi Zhang, Guilin Qi, Miaozeng Du, Yongrui Chen, Sheng Bi, and Fan Liu. Sin- gle image unlearning: Efficient machine unlearning in multi- modal large language models.Advances in Neural Informa- tion Processing Systems, 37:35414–35453, 2024. 2

    2024

  7. [15]

    Continual learning and private unlearning

    Bo Liu, Qiang Liu, and Peter Stone. Continual learning and private unlearning. InConference on Lifelong Learning Agents, pages 243–254. PMLR, 2022. 6, 11

    2022

  8. [16]

    Protecting privacy in multimodal large language models with mllmu- bench.arXiv preprint arXiv:2410.22108, 2024

    Zheyuan Liu, Guangyao Dou, Mengzhao Jia, Zhaoxuan Tan, Qingkai Zeng, Yongle Yuan, and Meng Jiang. Protecting privacy in multimodal large language models with mllmu- bench.arXiv preprint arXiv:2410.22108, 2024. 2, 7, 11

    2024 arXiv

  9. [17]

    Machine unlearning in generative ai: A survey.arXiv preprint arXiv:2407.20516, 2024

    Zheyuan Liu, Guangyao Dou, Zhaoxuan Tan, Yijun Tian, and Meng Jiang. Machine unlearning in generative ai: A survey.arXiv preprint arXiv:2407.20516, 2024. 1

    2024 arXiv

  10. [18]

    Towards safer large language models through machine unlearning.arXiv preprint arXiv:2402.10058, 2024

    Zheyuan Liu, Guangyao Dou, Zhaoxuan Tan, Yijun Tian, and Meng Jiang. Towards safer large language models through machine unlearning.arXiv preprint arXiv:2402.10058, 2024. 2

    2024 arXiv

  11. [19]

    Modality-aware neuron pruning for unlearning in multimodal large language models.arXiv preprint arXiv:2502.15910, 2025

    Zheyuan Liu, Guangyao Dou, Xiangchi Yuan, Chunhui Zhang, Zhaoxuan Tan, and Meng Jiang. Modality-aware neuron pruning for unlearning in multimodal large language models.arXiv preprint arXiv:2502.15910, 2025. 1, 2, 6, 12

    2025 arXiv

  12. [20]

    Tofu: A task of fictitious unlearn- ing for llms.arXiv preprint arXiv:2401.06121, 2024

    Pratyush Maini, Zhili Feng, Avi Schwarzschild, Zachary C Lipton, and J Zico Kolter. Tofu: A task of fictitious unlearn- ing for llms.arXiv preprint arXiv:2401.06121, 2024. 1, 2, 3, 6, 13

    2024 arXiv

  13. [21]

    The eu proposal for a general data protection regulation and the roots of the ‘right to be for- gotten’.Computer Law & Security Review, 29(3):229–235,

    Alessandro Mantelero. The eu proposal for a general data protection regulation and the roots of the ‘right to be for- gotten’.Computer Law & Security Review, 29(3):229–235,

  14. [22]

    Merging models with fisher-weighted averaging.Advances in Neural Information Processing Systems, 35:17703–17716, 2022

    Michael S Matena and Colin A Raffel. Merging models with fisher-weighted averaging.Advances in Neural Information Processing Systems, 35:17703–17716, 2022. 3

    2022

  15. [23]

    Eric Mitchell, Charles Lin, Antoine Bosselut, Chelsea Finn, and Christopher D. Manning. Fast model editing at scale. InInternational Conference on Learning Representations (ICLR), 2022. 3

    2022

  16. [24]

    Manning, and Chelsea Finn

    Eric Mitchell, Charles Lin, Antoine Bosselut, Christopher D. Manning, and Chelsea Finn. Memory-based model editing at scale. InInternational Conference on Machine Learning (ICML), pages 15817–15831. PMLR, 2022. 3 9

    2022

  17. [25]

    Variational bayesian unlearning.Advances in Neural Information Processing Systems, 33:16025–16036, 2020

    Quoc Phong Nguyen, Bryan Kian Hsiang Low, and Patrick Jaillet. Variational bayesian unlearning.Advances in Neural Information Processing Systems, 33:16025–16036, 2020. 2, 6, 11

    2020

  18. [26]

    Learn- ing transferable visual models from natural language super- vision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learn- ing transferable visual models from natural language super- vision. InICML, 2021. 1

    2021

  19. [27]

    Knowledge unlearning for llms: Tasks, methods, and challenges.arXiv preprint arXiv:2311.15766, 2023

    Nianwen Si, Hao Zhang, Heyu Chang, Wenlin Zhang, Dan Qu, and Weiqiang Zhang. Knowledge unlearning for llms: Tasks, methods, and challenges.arXiv preprint arXiv:2311.15766, 2023. 1, 3

    2023 arXiv

  20. [28]

    Unrolling sgd: Understanding factors in- fluencing machine unlearning

    Anvith Thudi, Gabriel Deza, Varun Chandrasekaran, and Nicolas Papernot. Unrolling sgd: Understanding factors in- fluencing machine unlearning. In2022 IEEE 7th European Symposium on Security and Privacy (EuroS&P), pages 303–

  21. [29]

    Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing in- ference time

    Mitchell Wortsman, Gabriel Ilharco, Samir Ya Gadre, Re- becca Roelofs, Raphael Gontijo-Lopes, Ari S Morcos, Hongseok Namkoong, Ali Farhadi, Yair Carmon, Simon Ko- rnblith, et al. Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing ...

    2022

  22. [30]

    Ties-merging: Resolving interference when merging models.Advances in Neural Information Pro- cessing Systems, 36:7093–7115, 2023

    Prateek Yadav, Derek Tam, Leshem Choshen, Colin A Raf- fel, and Mohit Bansal. Ties-merging: Resolving interference when merging models.Advances in Neural Information Pro- cessing Systems, 36:7093–7115, 2023. 3

    2023

  23. [31]

    Large language model unlearning.arXiv preprint arXiv:2310.10683, 2023

    Yuanshun Yao, Xiaojun Xu, and Yang Liu. Large language model unlearning.arXiv preprint arXiv:2310.10683, 2023. 2

    2023 arXiv

  24. [32]

    A survey on multimodal large language models.arXiv preprint arXiv:2306.13549, 2023

    Shukang Yin, Chaoyou Fu, Sirui Zhao, Ke Li, Xing Sun, Tong Xu, and Enhong Chen. A survey on multimodal large language models.arXiv preprint arXiv:2306.13549, 2023. 1

    2023 arXiv

  25. [33]

    Visual oriented fine-grained knowledge editing for multimodal large language models

    Zhen Zeng, Leijiang Gu, Xun Yang, Zhangling Duan, Zenglin Shi, and Meng Wang. Visual oriented fine-grained knowledge editing for multimodal large language models. In Proceedings of the International Conference on Computer Vision (ICCV), 2025. 3

    2025

  26. [34]

    A comprehensive study of knowledge editing for large language models.arXiv preprint arXiv:2401.01286, 2024

    Ningyu Zhang, Yunzhi Yao, Bozhong Tian, Peng Wang, Shumin Deng, Mengru Wang, Zekun Xi, Shengyu Mao, Jin- tian Zhang, Yuansheng Ni, et al. A comprehensive study of knowledge editing for large language models.arXiv preprint arXiv:2401.01286, 2024. 3

    2024 arXiv

  27. [35]

    I don’t know

    Ce Zheng, Lei Li, Qingxiu Dong, Yuxuan Fan, Zhiyong Wu, Jingjing Xu, and Baobao Chang. Can we edit factual knowl- edge by in-context learning? InEMNLP, 2023. 3 10 A. Appendix: Details of S-MLLMUn Bench We construct 1,000 profiles using Qwen-VL-Plus, with the detailed structure...

    2023

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.