Pith. sign in

REVIEW 4 major objections 5 minor 127 references

Erasing Concepts, Steering Generations: A Comprehensive Survey of Concept Suppression

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read This survey proposes a three-dimensional taxonomy—intervention level, optimization strategy, and semantic scope—that organizes concept erasure methods in text-to-image diffusion models so practitioners can compare trade-offs between…

desk verdict Useful survey with a solid taxonomy, but the central comparison tables are internally inconsistent and need major fixing. read the letter →

arxiv 2505.19398 v2 pith:6E66QACS submitted 2025-05-26 cs.CV

Yiwei Xie , Ping Liu , Zheng Zhang This is my paper
classification cs.CV
keywords concepterasuretext-to-imagediffusionmachineunlearningsafetyalignmentNSFWcontentcross-attentionclosed-formplug-inadapters
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

This survey paper claims that the scattered literature on concept erasure in text-to-image diffusion models can be organized along three independent axes: where in the model the intervention happens (text encoder, cross-attention, or UNet), how the suppression is learned (loss-based optimization, closed-form projections, plug-in adapters, or adversarial training), and what semantic scope the target has (a single explicit concept, multiple concepts, or the combination of individually benign concepts). It argues that this three-dimensional framing makes methodological comparisons clearer and exposes a recurring trade-off between erasure strength and model utility: deeper interventions suppress more reliably but cost more in fidelity and complexity. The survey further collects the evaluation metrics, datasets, and benchmarks used to measure erasure success and utility, and identifies open problems such as conceptual entanglement, reactivation attacks, and adaptation to new model architectures. A careful reader would care because the taxonomy is intended as a practical map for choosing or designing an erasure method for a specific safety requirement.

What carries the argument

The load-bearing object is the taxonomy itself, presented in Figure 2 as a three-axis scheme. The first axis, intervention level, splits methods into text-encoder-level (prompt embedding modifications), cross-attention-level (attention-map or key/value manipulation), and UNet-level (feature masking, pruning, loss-driven). The second axis, optimization structure, distinguishes loss-based objectives, closed-form projections (e.g., least-squares updates of attention projections), plug-in adapters (LoRA and gated modules), and adversarial training. The third axis, semantic scope, ranges from single explicit concepts to multi-concept and concept-combination erasure. The mechanism the taxonomy is supposed to carry is comparison: by locating any method in this space, a reader can see which architectural site, learning strategy, and target complexity it uses, and the two qualitative tables translate that location into expected effectiveness and utility.

What would settle it

A standardized evaluation that runs one representative method per taxonomy cell on the same prompts, metrics, and random seeds would falsify the survey's trade-off claims if the measured Effectiveness and Utility rankings contradict the ratings in Tables I and II (for example, if a text-encoder-level loss-based method outperformed every UNet-level method in both effectiveness and utility).

Watch

Extended reading notes

Core claim

On its own terms, the paper's central contribution is a three-dimensional classification framework for concept erasure techniques, with every reviewed method located by intervention level (text encoder, cross-attention, UNet), optimization strategy (loss-based, closed-form, plug-in adapters, adversarial training), and semantic scope (explicit, multi-concept, concept combination). It further claims that this organization reveals systematic trade-offs—for instance, that adversarial training gives the highest erasure effectiveness at the cost of low utility, while plug-in adapters preserve utility at moderate effectiveness—and that these trade-offs are consistent across intervention levels. The survey also asserts that current evaluation practice, centered on Erasure Success Rate, FID, and CLIP Score plus a growing set of benchmarks, is not yet standardized enough to support fully rigorous comparison, especially for robustness and practical deployment.

Load-bearing premise

The comparative conclusions rest on the paper's own qualitative Effectiveness and Utility ratings in Tables I and II, which are not backed by a common evaluation protocol or quantitative results.

Editorial extensions

If this is right

  • Given any candidate erasure method, a practitioner can locate it in the three axes and infer its likely trade-off profile, because the paper claims the taxonomy 'allows for clearer methodological comparisons.'
  • Methods that intervene more deeply (UNet-level) are predicted to achieve stronger, more specific suppression but with lower utility and higher computational cost, while shallow interventions trade robustness for efficiency and transferability.
  • Adversarial training consistently shows the highest effectiveness ratings but the lowest utility ratings across all intervention levels, so robust erasure should be paired with utility-preserving components.
  • No single method combines top erasure strength with top utility; the paper therefore argues that hybrid multi-layer, multi-objective pipelines are the promising next step.
  • The evaluation landscape is fragmented enough that the paper's recommended benchmarks and metrics (ESR, FID, CLIP Score, plus integrated datasets like UnlearnCanvas and HUB) will need further standardization before cross-method comparisons become truly reliable.

Reading between the lines

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

  • If the taxonomy is taken as a generative map, one testable extension is to use it to predict the performance of an unseen method from its coordinates, which the survey itself does not explicitly propose.
  • The paper rates effectiveness and utility without a shared evaluation protocol; a natural next step would be a standardized benchmark that computes ESR, FID, and CLIP Score under identical prompts and seeds for one representative method per cell of the taxonomy.
  • The framework's intervention-level axis is tied to the UNet/cross-attention architecture of Stable Diffusion; extending the survey to flow-matching models such as Flux, which the paper notes lack cross-attention, may require a fourth axis or a redefinition of intervention sites.
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. This survey reviews concept-erasure methods for text-to-image diffusion models, organizing the literature along three axes: intervention level (text encoder, cross-attention, UNet), optimization strategy (loss-based, closed-form, plug-in adapters, adversarial), and semantic scope (single, multi, compositional). It provides background definitions, a taxonomy diagram, equations for representative methods, qualitative effectiveness/utility ratings in Tables I and II, an overview of datasets and metrics, and discussion of challenges and future directions. The stated contribution is a structured framework enabling 'clearer methodological comparisons' and 'actionable insights' into erasure-utility trade-offs.

Significance. The survey addresses a rapidly growing and practically important area, and its temporal coverage through May 2025 is a genuine asset. The proposed taxonomy is plausible and largely consistent with the cited methods, and the compilation of datasets, metrics, and benchmark frameworks in Section IV is useful for practitioners. The paper does not present new experimental results, so its value rests on the accuracy, completeness, and internal consistency of its qualitative synthesis. If the taxonomy and ratings were made reliable, the survey would serve as a useful entry point for method selection; at present, the central comparative synthesis is undermined by unsupported and internally contradictory ratings in Tables I and II, and by several citation and notation errors that reduce confidence in the survey's accuracy.

major comments (4)
  1. [Section III.C, Tables I and II] These two tables are presented as transposed views of the same rating matrix, but several cells disagree. Text Encoder + Loss-based is rated Utility 'Moderate' in Table I and 'High' in Table II; Text Encoder + Closed-form is 'Moderate' in Table I and 'High' in Table II; Text Encoder + Adversarial is Effectiveness 'High' and Utility 'Low-Moderate' in Table I but 'Moderate' and 'Moderate' in Table II; UNet + Loss-based is 'Moderate' and 'Moderate' in Table I but 'High' and 'Low-Moderate' in Table II; Cross-Attention + Plug-in Adapters is Utility 'Moderate-High' in Table I and 'High' in Table II. Since these tables are the only support for the trade-off conclusions in Section III.C.d, the contradictions are load-bearing. The authors must either harmonize the tables and specify the protocol or source for each rating, including the mapping from reported metrics such as ESR, FID, and robustness to the ordinal labels, or explicitly label the ratings as subjective editorial opinion and temper the claims that rest on them.
  2. [Section III.C and Figure 2] The paper's central claim is a three-dimensional taxonomy, but Tables I and II only compare intervention level and optimization strategy; the third axis, semantic scope, is absent from the comparative synthesis. Section III.C.c discusses semantic scope in prose without any systematic comparison, so the claimed multidimensional comparison is only partially realized. Please either extend the comparison to include semantic scope or revise the claim to describe a two-dimensional comparison with semantic scope as a descriptive category rather than a compared dimension.
  3. [Section III.A.1.a, Eq. (7)] The update rule for Buster is written as y^{(n+1)} = -y^{(n)} - eta * grad E(y^{(n)}) + epsilon^{(n)}. With eta > 0, this is not a gradient-descent step and is inconsistent with the described objective of minimizing E; as written, the iteration diverges rather than converging. Please correct the equation or clarify the notation so that the description matches the cited method.
  4. [Figure 2, Section IV.A, and reference list] Several cited references do not correspond to the methods they are attached to. Figure 2 lists 'CA [97]' as an explicit-concept method, but reference [97] is the CAT cross-attention paper, while the CA method is reference [37]. Section IV.A identifies the 'Safety Checker' with reference [24] (NudeNet), whereas the Stable Diffusion safety checker is a different component typically associated with the model description in reference [99]. In addition, reference [74] is explicitly marked in the bibliography as withdrawn from ICLR 2025, yet it is presented in Section III.A.2.b without qualification. These inaccuracies undercut the survey's promise of being a reliable comparative guide and should be corrected.
minor comments (5)
  1. [Figure 2] The figure contains typos: 'Meta-Unlearing' should read 'Meta-Unlearning', 'Safety Checher' should read 'Safety Checker', and 'ErasingAnithing' should read 'EraseAnything'.
  2. [Section III.A.1.a] Near Eq. (7), 'Guassian' should be 'Gaussian'.
  3. [Section IV] The paper would benefit from a short paragraph describing the inclusion criteria and search protocol used to select the reviewed methods, since the claim of comprehensiveness is central to a survey's usefulness.
  4. [Figure 2 and Section III.A.2.a] Two distinct methods named 'ACE' ([52] and [66]) appear without disambiguation; please add descriptors such as 'Anti-editing Concept Erasure' and 'Attentional Concept Erasure' to avoid confusion.
  5. [Section III.A.1.e] In Eq. (14), the PromptSlider objective does not explicitly indicate the distribution over timestep t in the expectation; specifying it would improve reproducibility of the presented formulation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the survey's taxonomy and comparisons rest on external literature, not on self-cited or internally derived results.

full rationale

This is a survey paper with no derived equations, fitted parameters, or predictive claims, so the main circularity patterns do not apply. The central contribution is a three-dimensional taxonomy (intervention level, optimization strategy, semantic scope) constructed by organizing externally published concept-erasure methods; the methods and their reported behaviors are cited from the literature rather than generated by the authors' own framework. The only self-citations are references [10] and [11] in the Introduction, where prior work by co-author Ping Liu is cited as examples of text-to-image generation systems; these citations are background illustrations and do not support the taxonomy, the comparative claims, or any benchmark conclusion. The manuscript's comparative insights rely on the qualitative Effectiveness and Utility ratings in Tables I and II, and those ratings are presented without a stated evaluation protocol or source mapping, which is a support/consistency weakness rather than a circularity: the ratings are not derived from the taxonomy by definition, nor do they feed back into the definitions of the categories. Additionally, the two tables are internally inconsistent for several cells (e.g., Text Encoder + Loss-based rated 'Moderate' Utility in Table I but 'High' in Table II; Text Encoder + Adversarial rated 'High'/'Low-Moderate' in Table I but 'Moderate'/'Moderate' in Table II), but that is a correctness and reproducibility concern, not an instance of the paper's premises being equivalent to its conclusions by construction. No self-citation chain is load-bearing, no uniqueness theorem is imported from the authors' prior work, and no ansatz is smuggled in via citation. The survey is self-contained as a literature synthesis and does not reduce to its own inputs, so the circularity score is 0.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The survey introduces no free parameters, no new entities, and no derived quantities. Its central claims rest on the field's background assumption of semantic disentanglement, on the accuracy of the authors' reading of cited work, and on the validity of their qualitative comparison tables.

assumptions (3)
  • domain assumption Semantic concepts are at least partially disentangled within T2I model representations.
    Section II.D states that concept erasure methods "are typically grounded on the assumption that semantic concepts are at least partially disentangled within the model's internal representations." The surveyed intervention strategies all depend to some degree on this premise.
  • domain assumption The cited primary papers are accurately represented and the taxonomy categories are faithful to those papers.
    The survey's synthesis and qualitative tables depend on accurate readings of dozens of external papers. The text misattributes the Stable Diffusion safety checker to NudeNet [24], which shows that this assumption can fail.
  • ad hoc to paper The qualitative effectiveness and utility ratings in Tables I and II are informative without a shared benchmark or cited experimental basis.
    Tables I and II assign Low, Moderate, and High values to each intervention-level and optimization-strategy combination, but the paper does not describe a method for deriving these ratings or point to source experiments. The survey's comparative insights rely on this unstated premise.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Erasing Concepts, Steering Generations: A Comprehensive Survey of Concept Suppression." pith.science (2026). https://pith.science/paper/6E66QACS

@misc{pith2026250519398,
  author       = {Pith},
  title        = {Pith review of: Erasing Concepts, Steering Generations: A Comprehensive Survey of Concept Suppression},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6E66QACS}},
  note         = {Machine review of arXiv:2505.19398}
}
read the original abstract

Text-to-Image (T2I) models have demonstrated impressive capabilities in generating high-quality and diverse visual content from natural language prompts. However, uncontrolled reproduction of sensitive, copyrighted, or harmful imagery poses serious ethical, legal, and safety challenges. To address these concerns, the concept erasure paradigm has emerged as a promising direction, enabling the selective removal of specific semantic concepts from generative models while preserving their overall utility. This survey provides a comprehensive overview and in-depth synthesis of concept erasure techniques in T2I diffusion models. We systematically categorize existing approaches along three key dimensions: intervention level, which identifies specific model components targeted for concept removal; optimization structure, referring to the algorithmic strategies employed to achieve suppression; and semantic scope, concerning the complexity and nature of the concepts addressed. This multi-dimensional taxonomy enables clear, structured comparisons across diverse methodologies, highlighting fundamental trade-offs between erasure specificity, generalization, and computational complexity. We further discuss current evaluation benchmarks, standardized metrics, and practical datasets, emphasizing gaps that limit comprehensive assessment, particularly regarding robustness and practical effectiveness. Finally, we outline major challenges and promising future directions, including disentanglement of concept representations, adaptive and incremental erasure strategies, adversarial robustness, and new generative architectures. This survey aims to guide researchers toward safer, more ethically aligned generative models, providing foundational knowledge and actionable recommendations to advance responsible development in generative AI.

Figures

Figures reproduced from arXiv: 2505.19398 by the authors.

Figure 1
Figure 1. Concept erasure using ESD. The top two rows show image generations from the original model (left) and the unlearned model (right) across three [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Taxonomy of concept erasure techniques in text-to-image generative models, categorized along three orthogonal axes: (1) intervention level (e.g., [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Intervention-level categorization and corresponding optimization strategies for concept erasure in diffusion models. Concept erasure can be applied at [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Illustration of the text-to-image generation pipelines for (a) Diffusion [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

127 extracted references · 50 canonical work pages

  1. [24]

    Nudenet: Neural nets for nudity detection and censoring,

    P. Bedapudi, “Nudenet: Neural nets for nudity detection and censoring,” 2022. [Online]. Available: https://nudenet.notai.tech/

    2022

  2. [74]

    Towards robust concept erasure in diffusion models: Unlearning identity, nudity and artistic styles,

    U. Maharana, A. S. Sharma, Y . Sinha, A. Mali, M. Kankanhalli, and M. Mandal, “Towards robust concept erasure in diffusion models: Unlearning identity, nudity and artistic styles,” 2024, withdrawn from ICLR 2025 (OpenReview submission). [Online]. Available: https://openreview.net/forum?id=Ox2A1WoKLm

    work page 2024

  3. [97]

    Cat: Cross attention in vision transformer,

    H. Lin, X. Cheng, X. Wu, and D. Shen, “Cat: Cross attention in vision transformer,” in IEEE International Conference on Multimedia and Expo, 2022

    work page 2022

  4. [37]

    Ablating concepts in text-to-image diffusion models,

    N. Kumari, B. Zhang, S.-Y . Wang, E. Shechtman, R. Zhang, and J.-Y . Zhu, “Ablating concepts in text-to-image diffusion models,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023

    2023

  5. [99]

    High-resolution image synthesis with latent diffusion models,

    R. Rombach, A. Blattmann, D. Lorenz, P. Esser, and B. Ommer, “High-resolution image synthesis with latent diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022

    work page 2022

  6. [1]

    Generative adversarial net- works,

    I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversarial net- works,” Commun. ACM, 2020

    2020

  7. [2]

    Triple generative adversarial networks,

    C. Li, K. Xu, J. Zhu, J. Liu, and B. Zhang, “Triple generative adversarial networks,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022

    2022

  8. [3]

    Edge guided gans with multi-scale contrastive learning for semantic image synthesis,

    H. Tang, G. Sun, N. Sebe, and L. Van Gool, “Edge guided gans with multi-scale contrastive learning for semantic image synthesis,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2023

    2023

Show all 127 references
  1. [4]

    Auto-encoding variational bayes,

    D. P. Kingma and M. Welling, “Auto-encoding variational bayes,” arXiv:1312.6114, 2022

    2022 arXiv

  2. [5]

    Qarv: Quantization-aware resnet vae for lossy image compression,

    Z. Duan, M. Lu, J. Ma, Y . Huang, Z. Ma, and F. Zhu, “Qarv: Quantization-aware resnet vae for lossy image compression,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2024

    2024

  3. [6]

    Controlvae: Tuning, analytical properties, and performance analysis,

    H. Shao, Z. Xiao, S. Yao, D. Sun, A. Zhang, S. Liu, T. Wang, J. Li, and T. Abdelzaher, “Controlvae: Tuning, analytical properties, and performance analysis,” IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022

    2022

  4. [7]

    Denoising diffusion probabilistic models,

    J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” in Proceedings of the 34th International Conference on Neural Information Processing Systems , 2020

    2020

  5. [8]

    Diffi2i: Efficient diffusion model for image-to-image translation,

    B. Xia, Y . Zhang, S. Wang, Y . Wang, X. Wu, Y . Tian, W. Yang, R. Timotfe, and L. Van Gool, “Diffi2i: Efficient diffusion model for image-to-image translation,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2025

    2025

  6. [9]

    Create your world: Lifelong text-to-image diffusion,

    G. Sun, W. Liang, J. Dong, J. Li, Z. Ding, and Y . Cong, “Create your world: Lifelong text-to-image diffusion,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2024

    2024

  7. [10]

    Text-guided eyeglasses manipula- tion with spatial constraints,

    J. Wang, P. Liu, J. Liu, and W. Xu, “Text-guided eyeglasses manipula- tion with spatial constraints,” IEEE Transactions on Multimedia, 2024

    2024

  8. [11]

    Unified diffusion-based rigid and non- rigid editing with text and image guidance,

    J. Wang, P. Liu, and W. Xu, “Unified diffusion-based rigid and non- rigid editing with text and image guidance,” in IEEE International Conference on Multimedia and Expo , 2024

    2024

  9. [12]

    Discrimina- tive probing and tuning for text-to-image generation,

    L. Qu, W. Wang, Y . Li, H. Zhang, L. Nie, and T.-S. Chua, “Discrimina- tive probing and tuning for text-to-image generation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, 2024

    2024

  10. [13]

    A comprehensive survey on concept erasure in text-to-image diffusion models,

    C. Kim and Y . Qi, “A comprehensive survey on concept erasure in text-to-image diffusion models,” arXiv:2502.14896, 2025

    2025 arXiv

  11. [14]

    Erasing concepts from diffusion models,

    R. Gandikota, J. Materzynska, J. Fiotto-Kaufman, and D. Bau, “Erasing concepts from diffusion models,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2023

    2023

  12. [15]

    Forget-me-not: Learning to forget in text-to-image diffusion models,

    G. Zhang, K. Wang, X. Xu, Z. Wang, and H. Shi, “Forget-me-not: Learning to forget in text-to-image diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition, 2024

    2024

  13. [16]

    Race: Robust adversarial concept erasure for secure text-to-image diffusion model,

    C. Kim, K. Min, and Y . Yang, “Race: Robust adversarial concept erasure for secure text-to-image diffusion model,” in Proceedings of the European Conference on Computer Vision , 2024

    2024

  14. [17]

    Fast yet effective machine unlearning,

    A. K. Tarun, V . S. Chundawat, M. Mandal, and M. Kankanhalli, “Fast yet effective machine unlearning,” IEEE Transactions on Neural Networks and Learning Systems , 2024

    2024

  15. [18]

    Zero- shot machine unlearning,

    V . S. Chundawat, A. K. Tarun, M. Mandal, and M. Kankanhalli, “Zero- shot machine unlearning,” IEEE Transactions on Information Forensics and Security, 2023. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 15

    2023

  16. [19]

    Defensive unlearning with adversarial training for robust concept erasure in diffusion models,

    Y . Zhang, X. Chen, J. Jia, Y . Zhang, C. Fan, J. Liu, M. Hong, K. Ding, and S. Liu, “Defensive unlearning with adversarial training for robust concept erasure in diffusion models,” in The Thirty-eighth Annual Conference on Neural Information Processing Systems , 2024

    2024

  17. [20]

    Machine unlearning: A survey,

    H. Xu, T. Zhu, L. Zhang, W. Zhou, and P. S. Yu, “Machine unlearning: A survey,” ACM Comput. Surv., 2023

    2023

  18. [21]

    Erasing concepts from text-to-image diffusion models with few-shot unlearning,

    M. Fuchi and T. Takagi, “Erasing concepts from text-to-image diffusion models with few-shot unlearning,” in 35th British Machine Vision Conference, 2024

    2024

  19. [22]

    Safetext: Safe text-to-image models via aligning the text encoder,

    Y . Hu, Z. Jiang, and N. Z. Gong, “Safetext: Safe text-to-image models via aligning the text encoder,” arXiv:2502.20623, 2025

    2025 arXiv

  20. [23]

    Buster: Implanting semantic backdoor into text encoder to mitigate nsfw content generation,

    X. Zhao, X. Chen, Y . Xuan, Z. Zhao, X. Jia, X. Li, and X. Wang, “Buster: Implanting semantic backdoor into text encoder to mitigate nsfw content generation,” arXiv:2412.07249, 2025

    2025 arXiv

  21. [25]

    Espresso: Robust concept filtering in text-to-image models,

    A. Das, V . Duddu, R. Zhang, and N. Asokan, “Espresso: Robust concept filtering in text-to-image models,” in Proceedings of the 15th ACM Conference on Data and Application Security and Privacy , 2025

    2025

  22. [26]

    Ce- sdwv: Effective and efficient concept erasure for text-to-image diffusion models via a semantic-driven word vocabulary,

    J. Tu, Q. Feng, C. Chen, J. Dong, H. Zhao, C. Zhang, and H. Qian, “Ce- sdwv: Effective and efficient concept erasure for text-to-image diffusion models via a semantic-driven word vocabulary,” arXiv:2501.15562, 2025

    2025

  23. [27]

    Safe and reliable diffusion models via subspace projection,

    H. Chen, T. Zhu, L. Wang, X. Yu, L. Gao, and W. Zhou, “Safe and reliable diffusion models via subspace projection,” arXiv:2503.16835, 2025

    2025 arXiv

  24. [28]

    Sparse autoencoder as a zero-shot classifier for concept erasing in text-to-image diffusion models,

    Z. Tian, S. Nan, M. Xu, S. Zhai, W. Qu, J. Liu, K. Ren, R. Jia, and J. Zhang, “Sparse autoencoder as a zero-shot classifier for concept erasing in text-to-image diffusion models,” arXiv:2503.09446, 2025

    2025 arXiv

  25. [29]

    Concept steerers: Leveraging k-sparse autoencoders for controllable generations,

    D. Kim and D. Ghadiyaram, “Concept steerers: Leveraging k-sparse autoencoders for controllable generations,” arXiv:2501.19066, 2025

    2025

  26. [30]

    Saeuron: Interpretable concept unlearning in diffusion models with sparse autoencoders,

    B. Cywi ´nski and K. Deja, “Saeuron: Interpretable concept unlearning in diffusion models with sparse autoencoders,” in Proceedings of the 42th International Conference on Machine Learning , 2025

    2025

  27. [31]

    Distorting embedding space for safety: A defense mechanism for adversarially robust diffusion models,

    J. Ahn and H. Jung, “Distorting embedding space for safety: A defense mechanism for adversarially robust diffusion models,” arXiv:2501.18877, 2025

    2025

  28. [32]

    Prompt sliders for fine-grained control, editing and erasing of concepts in diffusion models,

    D. Sridhar and N. Vasconcelos, “Prompt sliders for fine-grained control, editing and erasing of concepts in diffusion models,” in Proceedings of the European Conference on Computer Vision , 2024

    2024

  29. [33]

    Promptguard: Soft prompt-guided unsafe content moderation for text-to-image models,

    L. Yuan, X. Jia, Y . Huang, W. Dong, and Y . Liu, “Promptguard: Soft prompt-guided unsafe content moderation for text-to-image models,” arXiv:2501.03544, 2025

    2025 arXiv

  30. [34]

    Safe text-to-image generation: Simply sanitize the prompt embedding,

    H. Qiu, G. Chen, M. Zhang, and M. Yang, “Safe text-to-image generation: Simply sanitize the prompt embedding,” arXiv:2411.10329, 2024

    2024 arXiv

  31. [35]

    Responsible diffusion models via constraining text embeddings within safe regions,

    Z. Li, D. Chen, M. Fan, C. Chen, Y . Li, Y . Wang, and W. Zhou, “Responsible diffusion models via constraining text embeddings within safe regions,” in Proceedings of the ACM on Web Conference , 2025

    2025

  32. [36]

    Crce: Coreference-retention concept erasure in text-to-image diffusion models,

    Y . Xue, E. Moroshko, F. Chen, S. McDonagh, and S. A. Tsaftaris, “Crce: Coreference-retention concept erasure in text-to-image diffusion models,” arXiv:2503.14232, 2025

    2025 arXiv

  33. [38]

    Choose your anchor wisely: Effective unlearning diffusion models via concept reconditioning,

    J. Zhu, R. Zhang, L. Lin, and S. Mei, “Choose your anchor wisely: Effective unlearning diffusion models via concept reconditioning,” in Proceedings of the Neural Information Processing Systems , 2024

    2024

  34. [39]

    Dark miner: Defend against undesired generation for text-to-image diffusion models,

    Z. Meng, B. Peng, X. Jin, Y . Jiang, J. Dong, and W. Wang, “Dark miner: Defend against undesired generation for text-to-image diffusion models,” arXiv:2409.17682, 2024

    2024 arXiv

  35. [40]

    Separable multi- concept erasure from diffusion models,

    M. Zhao, L. Zhang, T. Zheng, Y . Kong, and B. Yin, “Separable multi- concept erasure from diffusion models,” arXiv:2402.05947, 2024

    2024 arXiv

  36. [41]

    Erasing concept combination from text-to-image diffusion model,

    hongyi nie, Q. Yao, Y . Liu, Z. Wang, and Y . Bian, “Erasing concept combination from text-to-image diffusion model,” in Proceedings of the Thirteenth International Conference on Learning Representations , 2025

    2025

  37. [42]

    Continual unlearning for foundational text-to-image models without generaliza- tion erosion,

    K. Thakral, T. Glaser, T. Hassner, M. Vatsa, and R. Singh, “Continual unlearning for foundational text-to-image models without generaliza- tion erosion,” arXiv:2503.13769, 2025

    2025 arXiv

  38. [43]

    All but one: Surgical concept erasing with model preservation in text-to-image diffusion models,

    S. Hong, J. Lee, and S. S. Woo, “All but one: Surgical concept erasing with model preservation in text-to-image diffusion models,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2024

    2024

  39. [44]

    Growth inhibitors for suppressing inappropriate image concepts in diffusion models,

    D. Chen, Z. Li, M. Fan, C. Chen, W. Zhou, Y . Wang, and Y . Li, “Growth inhibitors for suppressing inappropriate image concepts in diffusion models,” in The Thirteenth International Conference on Learning Representations, 2025

    2025

  40. [45]

    Concept pinpoint eraser for text-to-image diffusion models via residual attention gate,

    B. H. Lee, S. Lim, S. Lee, D. U. Kang, and S. Y . Chun, “Concept pinpoint eraser for text-to-image diffusion models via residual attention gate,” in Proceedings of the Thirteenth International Conference on Learning Representations, 2025

    2025

  41. [46]

    Concept replacer: Replacing sensitive concepts in diffusion models via precision localization,

    L. Zhang, Y . Xie, Y . Fu, and P. Chen, “Concept replacer: Replacing sensitive concepts in diffusion models via precision localization,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025

    2025

  42. [47]

    Detect-and-guide: Self- regulation of diffusion models for safe text-to-image generation via guideline token optimization,

    F. Li, M. Zhang, Y . Sun, and M. Yang, “Detect-and-guide: Self- regulation of diffusion models for safe text-to-image generation via guideline token optimization,” in Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition , 2025

    2025

  43. [48]

    Casteer: Steering diffusion models for controllable generation,

    T. Gaintseva, C. Ma, Z. Liu, M. Benning, G. Slabaugh, J. Deng, and I. Elezi, “Casteer: Steering diffusion models for controllable generation,” arXiv:2503.09630, 2025

    2025 arXiv

  44. [49]

    Unified concept editing in diffusion models,

    R. Gandikota, H. Orgad, Y . Belinkov, J. Materzy ´nska, and D. Bau, “Unified concept editing in diffusion models,” in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision , 2024

    2024

  45. [50]

    Reliable and efficient concept erasure of text-to-image diffusion models,

    C. Gong, K. Chen, Z. Wei, J. Chen, and Y .-G. Jiang, “Reliable and efficient concept erasure of text-to-image diffusion models,” in Proceedings of the European Conference on Computer Vision , 2024

    2024

  46. [51]

    Mace: Mass concept erasure in diffusion models,

    S. Lu, Z. Wang, L. Li, Y . Liu, and A. W.-K. Kong, “Mace: Mass concept erasure in diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024

    2024

  47. [52]

    Ace: Anti- editing concept erasure in text-to-image models,

    Z. Wang, Y . Wei, F. Li, R. Pei, H. Xu, and W. Zuo, “Ace: Anti- editing concept erasure in text-to-image models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2025

    2025

  48. [53]

    Realera: Semantic-level concept erasure via neighbor-concept min- ing,

    Y . Liu, J. An, W. Zhang, M. Li, D. Wu, J. Gu, Z. Lin, and W. Wang, “Realera: Semantic-level concept erasure via neighbor-concept min- ing,” arXiv:2410.09140, 2024

    2024 arXiv

  49. [54]

    Trce: Towards reliable malicious concept erasure in text-to-image diffusion models,

    R. Chen, H. Guo, L. Wang, C. Zhang, W. Nie, and A.-A. Liu, “Trce: Towards reliable malicious concept erasure in text-to-image diffusion models,” arXiv:2503.07389, 2025

    2025

  50. [55]

    Precise, fast, and low-cost concept erasure in value space: Orthogonal complement matters,

    Y . Wang, O. Li, T. Mu, Y . Hao, K. Liu, X. Wang, and X. He, “Precise, fast, and low-cost concept erasure in value space: Orthogonal complement matters,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2025

    2025

  51. [56]

    Speed: Scalable, precise, and efficient concept erasure for diffusion models,

    O. Li, Y . Wang, X. Hu, H. Jiang, T. Liang, Y . Hao, G. Ma, and F. Feng, “Speed: Scalable, precise, and efficient concept erasure for diffusion models,” arXiv:2503.07392, 2025

    2025

  52. [57]

    Hiding and recovering knowledge in text-to-image diffusion models via learnable prompts,

    A. T. Bui, K. Doan, T. Le, P. Montague, T. Abraham, and D. Phung, “Hiding and recovering knowledge in text-to-image diffusion models via learnable prompts,” in The Thirteenth International Conference on Learning Representations, 2025

    2025

  53. [58]

    Direct unlearning optimization for robust and safe text-to- image models,

    Y .-H. Park, S. Yun, J.-H. Kim, J. Kim, G. Jang, Y . Jeong, J. Jo, and G. Lee, “Direct unlearning optimization for robust and safe text-to- image models,” in The Thirty-eighth Annual Conference on Neural Information Processing Systems , 2024

    2024

  54. [59]

    Conceptprune: Concept editing in diffusion models via skilled neuron pruning,

    R. Chavhan, D. Li, and T. Hospedales, “Conceptprune: Concept editing in diffusion models via skilled neuron pruning,” in The Thirteenth International Conference on Learning Representations , 2025

    2025

  55. [60]

    Pruning for robust concept erasing in diffusion models,

    T. Yang, Z. Li, J. Cao, and C. Xu, “Pruning for robust concept erasing in diffusion models,” in Neurips Safe Generative AI Workshop , 2024

    2024

  56. [61]

    Unveiling concept attribution in diffusion models,

    Q. H. Nguyen, H. Phan, and K. D. Doan, “Unveiling concept attribution in diffusion models,” arXiv:2412.02542, 2025

    2025

  57. [62]

    Robust concept erasure using task vectors,

    M. Pham, K. O. Marshall, C. Hegde, and N. Cohen, “Robust concept erasure using task vectors,” arXiv:2404.03631, 2025

    2025 arXiv

  58. [63]

    Localized concept erasure for text-to-image diffusion models using training-free gated low-rank adaptation,

    B. H. Lee, S. Lim, and S. Y . Chun, “Localized concept erasure for text-to-image diffusion models using training-free gated low-rank adaptation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2025

    2025

  59. [64]

    Receler: Reliable concept erasing of text-to-image diffu- sion models via lightweight erasers,

    C.-P. Huang, K.-P. Chang, C.-T. Tsai, Y .-H. Lai, F.-E. Yang, and Y .- C. F. Wang, “Receler: Reliable concept erasing of text-to-image diffu- sion models via lightweight erasers,” in Proceedings of the European Conference on Computer Vision , 2024

    2024

  60. [65]

    One-dimensional adapter to rule them all: Concepts, diffusion models and erasing applications,

    M. Lyu, Y . Yang, H. Hong, H. Chen, X. Jin, Y . He, H. Xue, J. Han, and G. Ding, “One-dimensional adapter to rule them all: Concepts, diffusion models and erasing applications,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024

    2024

  61. [66]

    Ace: Attentional concept erasure in diffusion models,

    F. Carter, “Ace: Attentional concept erasure in diffusion models,” arXiv:2504.11850, 2025. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 16

    2025 arXiv

  62. [67]

    Erasing undesirable influence in diffusion models,

    J. Wu, T. Le, M. Hayat, and M. Harandi, “Erasing undesirable influence in diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2025

    2025

  63. [68]

    Advan- chor: Enhancing diffusion model unlearning with adversarial anchors,

    M. Zhao, L. Zhang, X. Yang, T. Zheng, and B. Yin, “Advan- chor: Enhancing diffusion model unlearning with adversarial anchors,” arXiv:2501.00054, 2024

    2024 arXiv

  64. [69]

    Continuous concepts removal in text-to-image diffusion models,

    T. Han, W. Sun, Y . Hu, C. Fang, Y . Zhang, S. Ma, T. Zheng, Z. Chen, and Z. Wang, “Continuous concepts removal in text-to-image diffusion models,” arXiv:2412.00580, 2025

    2025 arXiv

  65. [70]

    Set you straight: Auto-steering denoising trajectories to sidestep unwanted concepts,

    L. Li, S. Lu, Y . Ren, and A. W.-K. Kong, “Set you straight: Auto-steering denoising trajectories to sidestep unwanted concepts,” arXiv:2504.12782, 2025

    2025 arXiv

  66. [71]

    Salun: Empowering machine unlearning via gradient-based weight saliency in both image classification and generation,

    C. Fan, J. Liu, Y . Zhang, E. Wong, D. Wei, and S. Liu, “Salun: Empowering machine unlearning via gradient-based weight saliency in both image classification and generation,” in Proceedings of the Twelfth International Conference on Learning Representatios , 2024

    2024

  67. [72]

    Erasing undesirable concepts in diffusion models with ad- versarial preservation,

    A. Bui, L. Vuong, K. Doan, T. Le, P. Montague, T. Abraham, and D. Phung, “Erasing undesirable concepts in diffusion models with ad- versarial preservation,” in Advances in Neural Information Processing Systems, 2024

    2024

  68. [73]

    Fantastic targets for concept erasure in diffusion models and where to find them,

    A. T. Bui, T.-T. Vu, L. T. Vuong, T. Le, P. Montague, T. Abraham, J. Kim, and D. Phung, “Fantastic targets for concept erasure in diffusion models and where to find them,” in The Thirteenth International Conference on Learning Representations , 2025

    2025

  69. [75]

    Unlearning concepts in diffusion model via concept domain correction and concept preserving gradient,

    Y . Wu, S. Zhou, M. Yang, L. Wang, H. Chang, W. Zhu, X. Hu, X. Zhou, and X. Yang, “Unlearning concepts in diffusion model via concept domain correction and concept preserving gradient,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2025

    2025

  70. [76]

    Stereo: Towards adversarially robust concept erasing from text-to-image gen- eration models,

    K. Srivatsan, F. Shamshad, M. Naseer, and K. Nandakumar, “Stereo: Towards adversarially robust concept erasing from text-to-image gen- eration models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2024

    2024

  71. [77]

    Efficient fine-tuning and concept suppression for pruned diffusion models,

    R. Shirkavand, P. Yu, S. Gao, G. Somepalli, T. Goldstein, and H. Huang, “Efficient fine-tuning and concept suppression for pruned diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2025

    2025

  72. [78]

    Meta-unlearning on diffusion models: Preventing relearning unlearned concepts,

    H. Gao, T. Pang, C. Du, T. Hu, Z. Deng, and M. Lin, “Meta-unlearning on diffusion models: Preventing relearning unlearned concepts,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025

    2025

  73. [79]

    Fine- grained erasure in text-to-image diffusion-based foundation models,

    K. Thakral, T. Glaser, T. Hassner, M. Vatsa, and R. Singh, “Fine- grained erasure in text-to-image diffusion-based foundation models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2025

    2025

  74. [80]

    Safe- gen: Mitigating sexually explicit content generation in text-to-image models,

    X. Li, Y . Yang, J. Deng, C. Yan, Y . Chen, X. Ji, and W. Xu, “Safe- gen: Mitigating sexually explicit content generation in text-to-image models,” in Proceedings of the 2024 on ACM SIGSAC Conference on Computer and Communications Security , 2024

    2024

  75. [81]

    Implicit concept removal of diffusion models,

    Z. Liu, K. Chen, Y . Zhang, J. Han, L. Hong, H. Xu, Z. Li, D.-Y . Yeung, and J. T. Kwok, “Implicit concept removal of diffusion models,” in Proceedings of the European Conference on Computer Vision , 2024

    2024

  76. [82]

    Dumo: Dual encoder modulation network for precise concept erasure,

    F. Han, K. Chen, C. Gong, Z. Wei, J. Chen, and Y .-G. Jiang, “Dumo: Dual encoder modulation network for precise concept erasure,” in Proceedings of the AAAI Conference on Artificial Intelligence , 2025

    2025

  77. [83]

    Concept corrector: Erase concepts on the fly for text-to-image diffusion models,

    Z. Meng, B. Peng, X. Jin, Y . Lyu, W. Wang, and J. Dong, “Concept corrector: Erase concepts on the fly for text-to-image diffusion models,” arXiv:2502.16368, 2025

    2025 arXiv

  78. [84]

    Safetydpo: Scalable safety alignment for text-to-image generation,

    R. Liu, C. I. Chieh, J. Gu, J. Zhang, R. Pi, Q. Chen, P. Torr, A. Khakzar, and F. Pizzati, “Safetydpo: Scalable safety alignment for text-to-image generation,” arXiv:2412.10493, 2024

    2024 arXiv

  79. [85]

    Crops: Model-agnostic training-free framework for safe image synthesis with latent diffusion models,

    J. Park, I. Ryu, J. Hwang, H. Park, J. Kim, and J.-S. Lee, “Crops: Model-agnostic training-free framework for safe image synthesis with latent diffusion models,” arXiv:2501.05359, 2025

    2025 arXiv

  80. [86]

    Munba: Machine unlearning via nash bargain- ing,

    J. Wu and M. Harandi, “Munba: Machine unlearning via nash bargain- ing,” arXiv:2411.15537, 2025

    2025 arXiv

  81. [87]

    Eraseanything: Enabling concept erasure in rectified flow transformers,

    D. Gao, S. Lu, S. Walters, W. Zhou, J. Chu, J. Zhang, B. Zhang, M. Jia, J. Zhao, Z. Fan, and W. Zhang, “Eraseanything: Enabling concept erasure in rectified flow transformers,” in Proceedings of the 42th International Conference on Machine Learning , 2025

    2025

  82. [88]

    Steerdiff: Steering towards safe text- to-image diffusion models,

    H. Zhang, Y . He, and H. Chen, “Steerdiff: Steering towards safe text- to-image diffusion models,” arXiv:2410.02710, 2024

    2024

  83. [89]

    Latent guard: a safety framework for text-to-image generation,

    R. Liu, A. Khakzar, J. Gu, Q. Chen, P. Torr, and F. Pizzati, “Latent guard: a safety framework for text-to-image generation,” inProceedings of the European Conference on Computer Vision , 2024

    2024

  84. [90]

    Guardt2i: Defending text-to-image models from adversarial prompts,

    Y . Yang, R. Gao, X. Yang, J. Zhong, and Q. Xu, “Guardt2i: Defending text-to-image models from adversarial prompts,” in The Thirty-eighth Annual Conference on Neural Information Processing Systems , 2024

    2024

  85. [91]

    Safe latent diffusion: Mitigating inappropriate degeneration in diffusion models,

    P. Schramowski, M. Brack, B. Deiseroth, and K. Kersting, “Safe latent diffusion: Mitigating inappropriate degeneration in diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023

    2023

  86. [92]

    Training-free safe denoisers for safe use of diffusion models,

    M. Kim, D. Kim, A. Yusuf, S. Ermon, and M. Park, “Training-free safe denoisers for safe use of diffusion models,” in Proceedings of the Thirteenth International Conference on Learning Representations , 2025

    2025

  87. [93]

    Trasce: Trajectory steering for concept erasure,

    A. Jain, Y . Kobayashi, T. Shibuya, Y . Takida, N. Memon, J. Togelius, and Y . Mitsufuji, “Trasce: Trajectory steering for concept erasure,” arXiv:2412.07658, 2025

    2025 arXiv

  88. [94]

    Safeguarding text- to-image generation via inference-time prompt-noise optimization,

    J. Peng, Z. Tang, G. Liu, C. Fleming, and M. Hong, “Safeguarding text- to-image generation via inference-time prompt-noise optimization,” arXiv:2412.03876, 2024

    2024 arXiv

  89. [95]

    Score forgetting distillation: A swift, data-free method for machine unlearning in diffusion models,

    T. Chen, S. Zhang, and M. Zhou, “Score forgetting distillation: A swift, data-free method for machine unlearning in diffusion models,” in The Thirteenth International Conference on Learning Representations , 2025

    2025

  90. [96]

    Data unlearning in diffusion models,

    S. Alberti, K. Hasanaliyev, M. Shah, and S. Ermon, “Data unlearning in diffusion models,” in Proceedings of the Thirteenth International Conference on Learning Representations , 2025

    2025

  91. [98]

    Learning transferable visual models from natural language supervi- sion,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark, G. Krueger, and I. Sutskever, “Learning transferable visual models from natural language supervi- sion,” in Proceedings of the 38th International Conference on Machine...

    2021

  92. [100]

    Red- teaming the stable diffusion safety filter,

    J. Rando, D. Paleka, D. Lindner, L. Heim, and F. Tramer, “Red- teaming the stable diffusion safety filter,” in Proceedings of the Neural Information Processing Systems , 2022

    2022

  93. [101]

    Nsfw detection machine learning model,

    G. Laborde, “Nsfw detection machine learning model,” 2020. [Online]. Available: https://github.com/GantMan/nsfw model

    2020

  94. [102]

    General data protection regulation (gdpr): Right to erasure (’right to be forgotten’),

    European Union, “General data protection regulation (gdpr): Right to erasure (’right to be forgotten’),” 2016. [Online]. Available: https://gdpr.eu/article-17-right-to-be-forgotten/

    2016

  95. [103]

    Recent advances in adversarial training for adversarial robustness,

    T. Bai, J. Luo, J. Zhao, B. Wen, and Q. Wang, “Recent advances in adversarial training for adversarial robustness,” in Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, 2021

    2021

  96. [104]

    GLIDE: Towards photorealistic image generation and editing with text-guided diffusion models,

    A. Q. Nichol, P. Dhariwal, A. Ramesh, P. Shyam, P. Mishkin, B. Mc- grew, I. Sutskever, and M. Chen, “GLIDE: Towards photorealistic image generation and editing with text-guided diffusion models,” in Proceedings of the 39th International Conference on Machine Learn- ing, 2022

    2022

  97. [105]

    Fastai: A layered api for deep learning,

    J. Howard and S. Gugger, “Fastai: A layered api for deep learning,” Information, 2020

    2020

  98. [106]

    Cifar-10 classification using deep convolutional neural network,

    R. Doon, T. Kumar Rawat, and S. Gautam, “Cifar-10 classification using deep convolutional neural network,” in IEEE Punecon, 2018

    2018

  99. [107]

    Large-scale classification of fine-art paint- ings: Learning the right metric on the right feature,

    B. Saleh and A. Elgammal, “Large-scale classification of fine-art paint- ings: Learning the right metric on the right feature,” arXiv:1505.00855, 2015

    2015 arXiv

  100. [108]

    Copyrightmeter: Revisiting copyright protection in text-to-image models,

    N. Xu, C. Li, T. Du, M. Li, W. Luo, J. Liang, Y . Li, X. Zhang, M. Han, J. Yin, and T. Wang, “Copyrightmeter: Revisiting copyright protection in text-to-image models,” arXiv:2411.13144, 2024

    2024 arXiv

  101. [109]

    Six-cd: Benchmarking concept removals for benign text-to- image diffusion models,

    J. Ren, K. Chen, Y . Cui, S. Zeng, H. Liu, Y . Xing, J. Tang, and L. Lyu, “Six-cd: Benchmarking concept removals for benign text-to- image diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2025

    2025

  102. [110]

    Holistic unlearning bench- mark: A multi-faceted evaluation for text-to-image diffusion model unlearning,

    S. Moon, M. Lee, S. Park, and D. Kim, “Holistic unlearning bench- mark: A multi-faceted evaluation for text-to-image diffusion model unlearning,” arXiv:2410.05664, 2025

    2025

  103. [111]

    Unlearncanvas: Stylized image dataset for enhanced machine unlearning evaluation in diffusion models,

    Y . Zhang, C. Fan, Y . Zhang, Y . Yao, J. Jia, J. Liu, G. Zhang, G. Liu, R. Kompella, X. Liu, and S. Liu, “Unlearncanvas: Stylized image dataset for enhanced machine unlearning evaluation in diffusion models,” in Proceedings of the Neural Information Processing Systems, 2024. ...

    2024

  104. [112]

    Microsoft coco: Common objects in context,

    T.-Y . Lin, M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan, P. Doll ´ar, and C. L. Zitnick, “Microsoft coco: Common objects in context,” in Proceedings of the European Conference on Computer Vision, 2014

    2014

  105. [113]

    Mma- diffusion: Multimodal attack on diffusion models,

    Y . Yang, R. Gao, X. Wang, T.-Y . Ho, N. Xu, and Q. Xu, “Mma- diffusion: Multimodal attack on diffusion models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2024

    2024

  106. [114]

    Ring-a-bell! how reliable are concept removal methods for diffusion models?

    Y .-L. Tsai, C.-Y . Hsu, C. Xie, C.-H. Lin, J. Y . Chen, B. Li, P.-Y . Chen, C.-M. Yu, and C.-Y . Huang, “Ring-a-bell! how reliable are concept removal methods for diffusion models?” in Proceedings of the Twelfth International Conference on Learning Representatios , 2024

    2024

  107. [115]

    Gans trained by a two time-scale update rule converge to a local nash equilibrium,

    M. Heusel, H. Ramsauer, T. Unterthiner, B. Nessler, and S. Hochreiter, “Gans trained by a two time-scale update rule converge to a local nash equilibrium,” in Proceedings of the Neural Information Processing Systems, 2017

    2017

  108. [116]

    CLIP- Score: A reference-free evaluation metric for image captioning,

    J. Hessel, A. Holtzman, M. Forbes, R. Le Bras, and Y . Choi, “CLIP- Score: A reference-free evaluation metric for image captioning,” in Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, 2021

    2021

  109. [117]

    T2isafety: Benchmark for assessing fairness, toxicity, and privacy in image generation,

    L. Li, Z. Shi, X. Hu, B. Dong, Y . Qin, X. Liu, L. Sheng, and J. Shao, “T2isafety: Benchmark for assessing fairness, toxicity, and privacy in image generation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2025

    2025

  110. [118]

    Erasebench: Understanding the ripple effects of concept erasure techniques,

    I. Amara, A. I. Humayun, I. Kajic, Z. Parekh, N. Harris, S. Young, C. Nagpal, N. Kim, J. He, C. N. Vasconcelos, D. Ramachandran, G. Farnadi, K. Heller, M. Havaei, and N. Rostamzadeh, “Erasebench: Understanding the ripple effects of concept erasure techniques,” arXiv:2501.09833, 2025

    2025

  111. [119]

    Erasing with precision: Evaluating specific concept erasure from text-to-image generative models,

    M. Fuchi and T. Takagi, “Erasing with precision: Evaluating specific concept erasure from text-to-image generative models,” arXiv:2502.13989, 2025

    2025 arXiv

  112. [120]

    Unified prompt attack against text-to-image generation models,

    D. Peng, Q. Ke, M. H. Huang, P. Hu, and J. Liu, “Unified prompt attack against text-to-image generation models,” IEEE Transactions on Pattern Analysis and Machine Intelligence , 2025

    2025

  113. [121]

    Prompting4debugging: red-teaming text-to-image diffusion models by finding problematic prompts,

    Z.-Y . Chin, C.-M. Jiang, C.-C. Huang, P.-Y . Chen, and W.-C. Chiu, “Prompting4debugging: red-teaming text-to-image diffusion models by finding problematic prompts,” in Proceedings of the 41st International Conference on Machine Learning , 2024

    2024

  114. [122]

    To generate or not? safety-driven unlearned diffusion models are still easy to generate unsafe images ... for now,

    Y . Zhang, J. Jia, X. Chen, A. Chen, Y . Zhang, J. Liu, K. Ding, and S. Liu, “To generate or not? safety-driven unlearned diffusion models are still easy to generate unsafe images ... for now,” in Proceedings of the European Conference on Computer Vision , 2024

    2024

  115. [123]

    Circumventing concept erasure methods for text-to-image generative models,

    M. Pham, K. O. Marshall, N. Cohen, G. Mittal, and C. Hegde, “Circumventing concept erasure methods for text-to-image generative models,” in Proceedings of the Twelfth International Conference on Learning Representatios, 2024

    2024

  116. [124]

    Devildiffusion: Embedding hidden noise backdoors into diffusion models,

    W. Aiken, P. Branco, and G.-V . Jourdan, “Devildiffusion: Embedding hidden noise backdoors into diffusion models,” in 2024 21st Annual International Conference on Privacy, Security and Trust (PST) , 2024

    2024

  117. [125]

    Eviledit: Backdooring text-to-image diffusion models in one second,

    H. Wang, S. Guo, J. He, K. Chen, S. Zhang, T. Zhang, and T. Xiang, “Eviledit: Backdooring text-to-image diffusion models in one second,” in Proceedings of the 32nd ACM International Conference on Multi- media, 2024

    2024

  118. [126]

    Rickrolling the artist: Injecting backdoors into text encoders for text-to-image synthesis,

    L. Struppek, D. Hintersdorf, and K. Kersting, “Rickrolling the artist: Injecting backdoors into text encoders for text-to-image synthesis,” in 2023 IEEE/CVF International Conference on Computer Vision (ICCV), 2023

    2023

  119. [127]

    Flow matching for generative modeling,

    Y . Lipman, R. T. Q. Chen, H. Ben-Hamu, M. Nickel, and M. Le, “Flow matching for generative modeling,” in Proceedings of the Eleventh International Conference on Learning Representations , 2023

    2023

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.