Pith. sign in

REVIEW 4 major objections 4 minor 30 references

DocMIA: Document-Level Membership Inference Attacks against DocVQA Models

T0 review · 4 major / 4 minor · reviewed 2026-08-09 · deepseek-v4-flash

Pith's one-line read A new membership-inference attack can tell whether a specific document was used to train a DocVQA model, with no auxiliary data.

desk verdict First document-level MIA for DocVQA, with a solid white-box story and a black-box evaluation that overstates transferability because the proxy is fit to the same documents it later classifies. read the letter →

arxiv 2502.03692 v1 pith:LDQ3GU6H submitted 2025-02-06 cs.LG cs.CLcs.CR

classification cs.LGcs.CLcs.CR
keywords membershipinferenceattackDocVQAdocument-levelprivacyoptimization-basedfeatureswhite-boxblack-boxdifferentialdefense
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 tries to establish that document-level membership inference against DocVQA models is practical without auxiliary data. It introduces DocMIA, which fine-tunes the target model (or a distilled proxy) on a single question-answer pair from a suspect document and measures how far the parameters move, how many steps convergence takes, and how the answer quality evolves. Aggregating these signals across a document's questions and clustering the result separates training documents from non-training documents. The authors report that these optimization-based features outperform existing membership-inference baselines across three DocVQA model families and two datasets, with balanced accuracy around 80% and F1 around 82% in the strongest cases. If correct, a white-box adversary, or a black-box adversary with query access and a proxy model, can audit or exploit whether confidential documents were in a DocVQA service's training data.

What carries the argument

The load-bearing object is the optimization-based distance feature $\Delta = \|\theta_0 - \theta_*\|_2$, the L2 norm of the parameter change when the target model is fine-tuned on one question-answer pair starting from its current weights. Together with the number of optimization steps $s$ and the aggregated DocVQA utility $u$, it forms a document-level descriptor after aggregation over the document's questions. The paper implements the fine-tuning in three variants: optimizing a single layer (FL), a LoRA adapter on that layer (FLLoRA), and the document image pixels (IG). In the black-box setting the machinery is a distillation step: the target model's answers on the suspect documents are used to train a public-checkpoint proxy, and the white-box features are extracted from the proxy to recover the membership signal.

What would settle it

Run the FL attack using only the $\Delta$ feature against a DocVQA model trained with strong regularization so that the train-test score gap is near zero and no document appears more than once; if balanced accuracy does not beat the average-score baseline, the claimed optimization-distance signal is not an independent membership signal.

Watch

Extended reading notes

Core claim

The central claim is that the optimization trajectory of a model on a single document-question-answer example is a membership signal. For a member document, fine-tuning the trained model on one of its question-answer pairs converges quickly and moves the parameters a small L2 distance, while a non-member document requires a larger parameter change; the number of steps and the DocVQA utility score supply supporting signals. The attack assigns each document a feature vector built from these quantities aggregated over its questions, then uses unsupervised clustering to label the cluster with the larger parameter distance as non-members. In the black-box setting the adversary first distills the target model's predictions into a proxy model and applies the same procedure to the proxy. The reported results show the strongest variants reaching roughly 80% balanced accuracy and 82% F1, exceeding loss-based, score-based, gradient-based, and token-probability baselines on the evaluated targets.

Load-bearing premise

The attack's success depends on the premise that the L2-norm of the parameter change after fine-tuning a single question-answer pair reliably separates member from non-member documents; the paper's own ablation shows that this distance alone underperforms a simple score on some targets, so the combined features and hyperparameters tuned on the evaluation set are what carry the result.

Editorial extensions

If this is right

  • A white-box adversary holding a suspect document can determine its training status with reported balanced accuracy up to about 80% and F1 up to about 82% on the evaluated models, without needing any auxiliary data.
  • A black-box adversary with API access can achieve similar leakage by distilling the target's answers into a proxy model; matching proxy and target architectures helps, but cross-architecture attacks still work in the reported experiments.
  • Documents with multiple question-answer pairs are exposed at the document level, so repeated exposure in training data amplifies membership leakage beyond single-example MIA.
  • The reported defense of training with DP-SGD at $\varepsilon = 8$ cuts the F1 of the strongest attack to about 55%, but at the cost of roughly halving DocVQA utility, so the defense is not free.
  • The attack remains partially effective when the adversary has only rephrased questions rather than the exact training questions, though all compared methods decline in that setting.

Reading between the lines

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

  • If parameter-space distance is the real leakage channel, then any training procedure that shortens the optimization path for held-out points, such as flatter minima or stronger regularization, should reduce membership-inference success; this is a testable defense target.
  • The same optimization-based recipe could transfer to other auto-regressive multimodal tasks such as general VQA or OCR-free document parsing, wherever per-sample fine-tuning is computationally feasible.
  • Because the authors tune hyperparameters per target model using the evaluation set, a realistic adversary without labeled member and non-member documents would likely see lower performance; evaluating with a fixed, untuned protocol would clarify the practical threat.
  • The black-box distillation result suggests the membership signal survives a prediction-only bottleneck; a useful extension would measure how much signal remains when the API returns only a top-1 answer without confidence scores.
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 / 4 minor

Summary. The paper introduces DocMIA, a document-level membership inference attack for DocVQA models. Two settings are considered: white-box attacks (FL, FLLoRA, IG) that fine-tune a single layer, LoRA parameters, or the document image and extract optimization-distance, step-count, and utility features; and a black-box attack that first distills a proxy model from the target model's outputs and then applies the white-box attack to the proxy. The authors report improved balanced accuracy and F1 over several score-, loss-, and gradient-based baselines across DocVQA and PFL-DocVQA and three model families, claiming state-of-the-art membership inference without auxiliary data.

Significance. If the empirical claims held, DocMIA would be a useful contribution: it is the first document-level MIA for DocVQA, it avoids shadow-training, and it addresses auto-regressive outputs in a structured way. The paper ships code, includes extensive ablations (hyperparameters, layer selection, feature ablation, rephrased questions, DP defenses), and evaluates across several models and datasets. These strengths make the work worth serious consideration. However, the evaluation protocol for the black-box setting has a transductive leak, and the white-box hyperparameters are tuned on the evaluation set; both issues undermine the strength of the reported numbers and the SOTA claim.

major comments (4)
  1. [§4.3, §5.2, Table 3] The black-box evaluation is transductive: Dquery is constructed from Dtest, the very same 600 documents later classified, and the proxy F_p is trained to maximize the likelihood of the target model's outputs on those documents. The proxy is therefore fit to the exact member and non-member documents it subsequently attacks; the optimization-distance feature Δ (Eq. 3) can simply encode how well the proxy memorized each document's target answers. Appendix F.3 and Figure 10(a) are consistent with this reading, as attack accuracy tracks the proxy's training loss approaching zero. To support the stated auxiliary-data-free claim, the proxy must be distilled on outputs for one set of documents and evaluated on a disjoint set of member/non-member documents.
  2. [Appendix C, Table 7] The hyperparameters α, τ, and layer L are selected by maximizing attack accuracy on Dtest, the same 600 documents used to report the headline numbers in Tables 2–4 and 11–16. Although the final clustering step is unsupervised, the hyperparameter search is supervised by the evaluation metric on the evaluation set. This makes the reported accuracies optimistic and the comparison to baselines (which use fixed rules, e.g., threshold at the average score) unfair. A validation split disjoint from the test documents should be used for hyperparameter selection, or the paper should report results with hyperparameters fixed a priori.
  3. [Table 3] The black-box DocMIA results do not consistently beat the baselines. For example, with the VT5 proxy on DVQA/VT5, FL gives −0.84 ACC and −1.10 F1 versus the best SCORE-UA baseline, and with the Donut proxy on DVQA/VT5, FL gives −4.34 ACC and −4.36 F1. The abstract's claim of outperforming state-of-the-art attacks across all models and datasets is therefore supported only for a subset of configurations (notably when the proxy architecture matches the target or when the target is Pix2Struct). Please report a clear win/loss count and a statistical comparison across seeds, or soften the claim.
  4. [Appendix F.1, Tables 14/15] The paper's central intuition is that the optimization-distance feature Δ is a reliable membership signal, but the feature-ablation tables show Δ alone is not discriminative on several targets: for Donut/DocVQA, AVG(Δ) gives F1 57.34 versus AVG(NLS) 76.88, and AVG(s) alone gives only 60.32. The large improvement from switching AVG to Φall (+9.24 F1 on Donut/DocVQA) indicates that the headline results are driven by feature combination and aggregation rather than by the optimization-distance premise per se. The paper should identify which features actually carry the signal and temper the narrative that optimization-based distance is the key novel ingredient.
minor comments (4)
  1. [Table 7] The header contains a typo: 'Hyperaremeters' should be 'Hyperparameters'.
  2. [Algorithm 1] The early-stopping condition on line 8, (L(θ) − l) < τ, is described in the text as stopping when 'no further improvements is observed,' but the condition as written stops when the loss decrease is below a threshold. Please clarify the intended sign and whether the check uses the current loss or the previous loss.
  3. [Figure 1] The caption says 'Training:' for the first panel, which is likely a typo for 'Train' or 'Training phase'; please adjust for clarity.
  4. [Appendix F.3] The discussion of the proxy model training to near-zero loss should be moved or summarized in the main text, since it directly affects the interpretation of the black-box results and the transductive concern raised above.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: DocMIA's reported gains are measured empirically against public checkpoints and fixed benchmarks; the black-box proxy design is transductive but not a definitional reduction.

full rationale

The manuscript is an empirical attack paper, not a derivation from first principles. The white-box distance feature Δ (Eq. 3) is an optimization statistic whose discriminative power is evaluated, not assumed; Appendix F.1 shows cases where Δ alone is weaker than the utility score (e.g., Donut/DocVQA F1 57.34 vs 76.88), so the method's success is not built into its own definition. The black-box construction in Sec. 4.3 trains the proxy on target outputs for Dtest and then attacks the same 600 documents (Sec. 5.3), which makes the evaluation transductive and raises a legitimate transferability concern: the proxy may have memorized target answers for exactly those documents, so the reported black-box gains may not transfer to a disjoint candidate set. However, this is not circularity under the required standard: the proxy is fit to target predictions, not to the member/non-member labels, and the attack score is a measured function of further fine-tuning, not algebraically equal to the proxy's training loss or to the membership label. Similarly, hyperparameters are tuned on the evaluation set (Sec. 5.2, App. C), which is an evaluation-integrity concern rather than a definitional reduction of the claimed result. Citations to Tito et al. (2024) anchor the public dataset, checkpoint, and one baseline; they are independent external anchors. No load-bearing step reduces by construction to its own input, so the circularity score is 0.

Assumptions & free parameters 7 free parameters · 4 assumptions · 0 invented entities

The paper does not invent new physical or mathematical entities. Its central claim rests on several domain assumptions: access to exact QA pairs, the reliability of fine-tuning distance as a membership signal, and the effectiveness of distillation transfer. The main free parameters are attack hyperparameters, several tuned on the evaluation set itself, which raises the burden on the empirical claims.

free parameters (7)
  • Early-stopping threshold tau for FL variant = VT5: 1e-6; Donut: 5.0; Pix2Struct-B: 1e-4 (Table 7)
    Controls when fine-tuning stops; selected per target model by maximizing attack accuracy on Dtest (Appendix C).
  • Early-stopping threshold tau for IG variant = VT5: 1e-5; Donut: not listed clearly; Pix2Struct-B: 1e-3 (Table 7)
    Same selection procedure as tau_FL, for the image-gradient variant.
  • Learning rate alpha for FL/FLLoRA = 0.001 for all models (Table 7)
    Chosen by grid search over {1e-4, 0.001, 0.01, 0.1, 0.5, 1.0} on the same Dtest (Appendix C).
  • Learning rate alpha for IG = 1.0 for VT5/Donut; 1e-4 for Pix2Struct-B (Table 7)
    Grid search over {0.001, 0.01, 0.1, 0.5, 1.0, 5.0, 10.0, 20.0}; tuned on Dtest.
  • Layer L to optimize = last decoder FC layer (embedding projection layer)
    Selected from the ablation in Table 6 as the layer giving highest attack accuracy.
  • Max optimization steps S = 200
    Fixed in all experiments to encourage convergence (Appendix C); not tuned.
  • Max questions per document M = 10
    Documents have varying question counts (Figure 6); the attack caps at 10 questions for tractability (Appendix C).
assumptions (4)
  • domain assumption The adversary knows the exact training question-answer pairs for each candidate document.
    Stated in Section 4.1 and used throughout Algorithm 1; the attack fine-tunes on (q_i, a_i) and aggregates over the document's exact questions. The appendix shows paraphrasing degrades all attacks.
  • domain assumption The fine-tuning distance and step count separate member from non-member documents.
    Core intuition in Section 4.2.1, not formally derived; the paper provides visualizations and ablations but the distance feature alone is weak for some targets (Table 14).
  • domain assumption In the black-box setting, a proxy trained on the target model's outputs for Dtest preserves membership-indicative structure.
    Section 4.3 assumes the transfer; validated empirically but not justified theoretically, and the proxy is trained on the documents later attacked.
  • domain assumption Public pre-trained checkpoints used for proxy initialization are available and are disjoint from the private training set.
    Section 4.3 states this; it makes the black-box attack not fully auxiliary-data-free in the strongest sense.

how reviews work

0 comments
Cite this review

Pith. "Pith review of DocMIA: Document-Level Membership Inference Attacks against DocVQA Models." pith.science (2026). https://pith.science/paper/LDQ3GU6H

@misc{pith2026250203692,
  author       = {Pith},
  title        = {Pith review of: DocMIA: Document-Level Membership Inference Attacks against DocVQA Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LDQ3GU6H}},
  note         = {Machine review of arXiv:2502.03692}
}
read the original abstract

Document Visual Question Answering (DocVQA) has introduced a new paradigm for end-to-end document understanding, and quickly became one of the standard benchmarks for multimodal LLMs. Automating document processing workflows, driven by DocVQA models, presents significant potential for many business sectors. However, documents tend to contain highly sensitive information, raising concerns about privacy risks associated with training such DocVQA models. One significant privacy vulnerability, exploited by the membership inference attack, is the possibility for an adversary to determine if a particular record was part of the model's training data. In this paper, we introduce two novel membership inference attacks tailored specifically to DocVQA models. These attacks are designed for two different adversarial scenarios: a white-box setting, where the attacker has full access to the model architecture and parameters, and a black-box setting, where only the model's outputs are available. Notably, our attacks assume the adversary lacks access to auxiliary datasets, which is more realistic in practice but also more challenging. Our unsupervised methods outperform existing state-of-the-art membership inference attacks across a variety of DocVQA models and datasets, demonstrating their effectiveness and highlighting the privacy risks in this domain.

Figures

Figures reproduced from arXiv: 2502.03692 by the authors.

Figure 1
Figure 1. The General Scheme of Document-level Membership Inference Attacks. Training: A DocVQA model is trained on a dataset of documents, each associated with multiple questions/answers. Deployment: An adversary exploits this structure by querying the model with questions related to a target document. By aggregating the model’s responses, the adversary can infer the membership of the document in the training set. of the doc… view at source ↗
Figure 2
Figure 2. Visualization of our fine-tuning strategy in the pa￾rameters space. Each contour plot represents the optimization landscape w.r.t each pair (ai, qi) from document x. In general, the average ∆ computed on a mem￾ber document xin is smaller than non-member document xout. Intuition. Since DocVQA models are typically trained on multiple question-answer pairs per document, the model parameters likely converge to minimize … view at source ↗
Figure 3
Figure 3. Variants of our proposed DocMIA. Left: (a) (b) illustrate three attack strategies in the white-box setting: optimizing either the Document Image or a Single Layer (LoRA). Dashed arrows indicate the back￾propagated gradient during optimization. Right: We distill the black-box into a proxy model, which is then attacked using the white-box strategies. with Kaiming initialization (He et al., 2015). From Algorithm 1, we … view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: White-box Setting: Our proposed at￾tacks consistently achieve high performance, gen￾erally outperforming the considered baselines. DVQA PFL VT5 Donut Pix2Struct-B VT5 Donut LOSS-TA 14.00 7.67 5.33 3.00 14.67 GRADIENT-UA 9.33 6.00 5.00 3.00 8.33 SCORELOSS-UAall 4.67 8.6…
Figure 5
Figure 5. Figure 5: MI performance versus the Train-Test gap. The target models exhibit varying Train-Test gaps, measured by the difference in DocVQA scores between member/non-member documents. Our attacks remain effective even when the gap is small, with performance improving as the gap …
Figure 6
Figure 6. Figure 6: The distribution of number per-document questions from PFL and DocVQA dataset. B BASELINES For the black-box setting, we evaluate three MI attacks as baselines, which only requires generated text to infer the membership of the target document: Score-Threshold Attack (S…
Figure 7
Figure 7. Figure 7: Ablation Study on Learning Rate α and Threshold τ . The best value for each model across all datasets is used as the hyperparameters in our black-box attacks. Layer VT5(PFL) Donut(DocVQA) Pix2Struct-B(DocVQA) Embedding Projection Layer 67.0 71.33 68.66 Embedding Layer …
Figure 8
Figure 8. Figure 8: Membership Features against three different target mod￾els on DocVQA Dataset. Top: The distribution of average loss over all questions from all target documents on each target model. Bottom: T￾SNE visualization of the features used in our proposed attacks. VT5 Pix2Stru…
Figure 10
Figure 10. Figure 10: The resulting Proxy Model against Pix2Struct-B in the black-box setting. (a) The attack accuracy improves quickly once the loss reaches near zero. (b) The optimization distance values between member and non-member documents exhibit a separation similar to that seen in…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

30 extracted references · 20 canonical work pages

  1. [1]

    Deep learning with differential privacy

    Martin Abadi, Andy Chu, Ian Goodfellow, H Brendan McMahan, Ilya Mironov, Kunal Talwar, and Li Zhang. Deep learning with differential privacy. In Proceedings of the 2016 ACM SIGSAC conference on computer and communications security, pp. 308–318,

  2. [5]

    Practical blind membership inference attack via differential comparisons

    Bo Hui, Yuchen Yang, Haolin Yuan, Philippe Burlina, Neil Zhenqiang Gong, and Yinzhi Cao. Practical blind membership inference attack via differential comparisons. arXiv preprint arXiv:2101.01341,

  3. [6]

    Our findings reveal thatlayers closer to the final output exhibit higher privacy leakagein terms of MI compared to (randomly selected) intermediate layers, likely due to receiving larger gradient updates. Specifically, fine-tuning the final fully connected layer alone leads to strong attack performance 17 Published as a conference paper at ICLR 2025 Model...

  4. [7]

    D M ORE ON ATTACK IMPLEMENTATION D.1 T ARGET MODEL TRAINING For all target models, whenever feasible, we utilize the public checkpoint fine-tuned on the consid- ered private dataset from Hugging Face library and adhere to the data processing guidelines, such as document resolution, as recommended by the authors. We deliberately opt for public checkpoints ...

  5. [8]

    ISBN 978-3-031-19815-1

    Springer Nature Switzerland. ISBN 978-3-031-19815-1. Diederik P Kingma. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980,

  6. [9]

    Pix2struct: screenshot pars- ing as pretraining for visual language understanding

    12 Published as a conference paper at ICLR 2025 Kenton Lee, Mandar Joshi, Iulia Turc, Hexiang Hu, Fangyu Liu, Julian Eisenschlos, Urvashi Khan- delwal, Peter Shaw, Ming-Wei Chang, and Kristina Toutanova. Pix2struct: screenshot pars- ing as pretraining for visual language understanding. In Proceedings of the 40th International Conference on Machine Learnin...

  7. [10]

    Membership leakage in label-only exposures

    Zheng Li and Yang Zhang. Membership leakage in label-only exposures. InProceedings of the 2021 ACM SIGSAC Conference on Computer and Communications Security, pp. 880–895,

  8. [12]

    Ml-leaks: Model and data independent membership inference attacks and defenses on machine learning models

    Ahmed Salem, Yang Zhang, Mathias Humbert, Pascal Berrang, Mario Fritz, and Michael Backes. Ml-leaks: Model and data independent membership inference attacks and defenses on machine learning models. arXiv preprint arXiv:1806.01246,

Show all 30 references
  1. [14]

    Membership inference at- tacks against machine learning models

    Reza Shokri, Marco Stronati, Congzheng Song, and Vitaly Shmatikov. Membership inference at- tacks against machine learning models. In 2017 IEEE symposium on security and privacy (SP), pp. 3–18. IEEE,

  2. [15]

    Privacy risks of securing machine learning models against adversarial examples

    Liwei Song, Reza Shokri, and Prateek Mittal. Privacy risks of securing machine learning models against adversarial examples. In Proceedings of the 2019 ACM SIGSAC conference on computer and communications security, pp. 241–257,

  3. [16]

    Privacy-aware document visual question answering

    13 Published as a conference paper at ICLR 2025 Rub`en Tito, Khanh Nguyen, Marlon Tobaben, Raouf Kerkouche, Mohamed Ali Souibgui, Kangsoo Jung, Joonas J ¨alk¨o, Vincent Poulain D ´Andecy, Aurelie Joseph, Lei Kang, et al. Privacy-aware document visual question answering. In Int...

  4. [17]

    Transformers: State-of-the-art natural language processing

    Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pierric Cistac, Tim Rault, Remi Louf, Morgan Funtowicz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mari...

  5. [18]

    Jiayuan Ye, Aadyaa Maddi, Sasi Kumar Murakonda, Vincent Bindschaedler, and Reza Shokri

    Association for Computational Linguistics. Jiayuan Ye, Aadyaa Maddi, Sasi Kumar Murakonda, Vincent Bindschaedler, and Reza Shokri. En- hanced membership inference attacks against machine learning models. In Proceedings of the 2022 ACM SIGSAC Conference on Computer and Communic...

  6. [19]

    Privacy risk in machine learn- ing: Analyzing the connection to overfitting

    Samuel Yeom, Irene Giacomelli, Matt Fredrikson, and Somesh Jha. Privacy risk in machine learn- ing: Analyzing the connection to overfitting. In 2018 IEEE 31st computer security foundations symposium (CSF), pp. 268–282. IEEE,

  7. [20]

    Min-k%++: Improved baseline for detecting pre-training data from large language models

    Jingyang Zhang, Jingwei Sun, Eric Yeats, Yang Ouyang, Martin Kuo, Jianyi Zhang, Hao Frank Yang, and Hai Li. Min-k%++: Improved baseline for detecting pre-training data from large language models. arXiv preprint arXiv:2404.02936,

  8. [22]

    This dataset is specifically designed for DocVQA tasks in a federated learning and differential privacy setup, supporting different levels of privacy granularity

    A large-scale dataset of real business invoices, often containing privacy-sensitive information such as payment amounts, tax numbers, and bank account details. This dataset is specifically designed for DocVQA tasks in a federated learning and differential privacy setup, suppor...

  9. [23]

    A document is predicted as a member if ¯l ≤ κ and otherwise non-member, where κ is selected as the average value of ¯l across Dtest

    Similar to S CORE -TA, this attack com- putes the average loss ¯l = 1 M ΣM i L(F (x, qi)). A document is predicted as a member if ¯l ≤ κ and otherwise non-member, where κ is selected as the average value of ¯l across Dtest. Unsupervised One-step Gradient Attack (GRADIENT -UA) ...

  10. [24]

    We first study the effect of α, which controls the speed of the optimization process in our attacks

    Learning Rate α. We first study the effect of α, which controls the speed of the optimization process in our attacks. This threshold τ is empirically set to be the average loss change observed when performing one optimization step after reaching the correct answer. Only the di...

  11. [27]

    Table 10 presents the target models’ performance across both DocVQA datasets

    D.2 T ARGET MODEL PERFORMANCE ON DOCVQA To ensure the utility of the target models for our experiments, we validated that the DocVQA per- formance of each model checkpoint closely matched the results reported in the respective papers. Table 10 presents the target models’ perfo...

  12. [28]

    total amount

    This performance can be attributed to the clear separation in the loss distribution between member and non-member samples (Figure 8), which indicates overfitting behavior in these cases. F M ORE ON ANALYSIS In this section, we provide a deeper analysis of the effectiveness of ...

  13. [29]

    The expectation is that the proxy model can capture internal decision-making patterns by following the black-box’s prediction strategies

    F.3 T HE RESULTING PROXY MODEL The purpose of training the Proxy Model on Dquery, with labels generated by the black-box model, is to mimic the prediction patterns of the black-box model. The expectation is that the proxy model can capture internal decision-making patterns by ...

  14. [1920]

    For additional details on the effects of document resolution, we refer readers to the original model’s paper(Kim et al., 2022)

    in the public checkpoint provided by the original authors, which slightly reduced the model’s DocVQA performance. For additional details on the effects of document resolution, we refer readers to the original model’s paper(Kim et al., 2022). The model was trained using the Ada...

  15. [2015]

    Lora: Low-rank adaptation of large language models

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685,

  16. [2018]

    Detecting pretraining data from large language models

    Weijia Shi, Anirudh Ajith, Mengzhou Xia, Yangsibo Huang, Daogao Liu, Terra Blevins, Danqi Chen, and Luke Zettlemoyer. Detecting pretraining data from large language models. arXiv preprint arXiv:2310.16789,

  17. [2019]

    Mem- bership inference attacks from first principles

    Nicholas Carlini, Steve Chien, Milad Nasr, Shuang Song, Andreas Terzis, and Florian Tramer. Mem- bership inference attacks from first principles. In 2022 IEEE Symposium on Security and Privacy (SP), pp. 1897–1914. IEEE,

  18. [2020]

    arXiv preprint arXiv:2008.08899,

  19. [2021]

    Mistral 7b

    Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, et al. Mistral 7b. arXiv preprint arXiv:2310.06825,

  20. [2022]

    Label-only membership inference attacks

    Christopher A Choquette-Choo, Florian Tramer, Nicholas Carlini, and Nicolas Papernot. Label-only membership inference attacks. In International conference on machine learning, pp. 1964–1974. PMLR,

  21. [2023]

    Ocr-free document un- derstanding transformer

    Geewook Kim, Teakgyu Hong, Moonbin Yim, JeongYeon Nam, Jinyoung Park, Jinyeong Yim, Wonseok Hwang, Sangdoo Yun, Dongyoon Han, and Seunghyun Park. Ocr-free document un- derstanding transformer. In Shai Avidan, Gabriel Brostow, Moustapha Ciss ´e, Giovanni Maria Farinella, and Ta...

  22. [2024]

    18 D.2 Target Model Performance on DocVQA

    14 Published as a conference paper at ICLR 2025 APPENDIX A DocVQA Datasets 15 B Baselines 16 C Ablation Study 17 D More on Attack Implementation 18 D.1 Target Model Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 D.2 Target Model Performance on DocV...

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.