mlsanityData to 7 Oct 2026

Membership Inference Attacks Against Fine-tuned Diffusion Language Models

arXiv 2601.20125 ↗First posted 27 Jan 2026cs.LG cs.AIcited by 5

On the map: Diffusion models → Robustness, privacy and adversarial aspects

also in 1 more
Abstract (arXiv)

Diffusion Language Models (DLMs) represent a promising alternative to autoregressive language models, using bidirectional masked token prediction. Yet their susceptibility to privacy leakage via Membership Inference Attacks (MIA) remains critically underexplored. This paper presents the first systematic investigation of MIA vulnerabilities in DLMs. Unlike the autoregressive models' single fixed prediction pattern, DLMs' multiple maskable configurations exponentially increase attack opportunities. This ability to probe many independent masks dramatically improves detection chances. To exploit this, we introduce SAMA (Subset-Aggregated Membership Attack), which addresses the sparse signal challenge through robust aggregation. SAMA samples masked subsets across progressive densities and applies sign-based statistics that remain effective despite heavy-tailed noise. Through inverse-weighted aggregation prioritizing sparse masks' cleaner signals, SAMA transforms sparse memorization detection into a robust voting mechanism. Experiments on nine datasets show SAMA achieves 30% relative AUC improvement over the best baseline, with up to 8 times improvement at low false positive rates. These findings reveal significant, previously unknown vulnerabilities in DLMs, necessitating the development of tailored privacy defenses.

Key facts. Cited by 5 papers in the map (data through 7 Oct 2026). Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models. Builds on 91 of its 109 references that are in the map. First posted on arXiv 27 Jan 2026.

Builds on

91 of 109 references.

Closest ancestors · 12

Same direction, newest first

  1. 2024Membership Inference on Text-to-Image Diffusion Models via Conditional Likelihood Discrepancy23 May
  2. 2024Privacy-Preserving Low-Rank Adaptation against Membership Inference Attacks for Latent Diffusion Models19 Feb
  3. 2023Black-box Membership Inference Attacks against Fine-tuned Diffusion Models13 Dec
  4. 2023Membership Inference Attacks on Diffusion Models via Quantile Regression8 Dec
  5. 2023Generated Distributions Are All You Need for Membership Inference Attacks Against Generative Models30 Oct
  6. 2023A Probabilistic Fluctuation based Membership Inference Attack for Diffusion Models23 Aug
  7. 2023White-box Membership Inference Attacks against Diffusion Models11 Aug
  8. 2023Towards More Realistic Membership Inference Attacks on Large Diffusion Models22 Jun
  9. 2023An Efficient Membership Inference Attack for the Diffusion Model by Proximal Initialization26 May
  10. 2023Membership Inference Attacks against Diffusion Models7 Feb
  11. 2023Are Diffusion Models Vulnerable to Membership Inference Attacks?2 Feb
  12. 2023Membership Inference of Diffusion Models24 Jan

Other references · 79

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  4. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  5. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,393
  6. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  7. 2017On Calibration of Modern Neural Networkscited by 4,860
  8. 2016Pointer Sentinel Mixture Modelscited by 2,992
  9. 2015Character-level Convolutional Networks for Text Classificationcited by 2,293
  10. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,221
  11. 2020Extracting Training Data from Large Language Modelscited by 2,144
  12. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,665
38 more
  1. 2018Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarizationcited by 1,308
  2. 2016Optimization Methods for Large-Scale Machine Learningcited by 1,140
  3. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,006
  4. 2025Large Language Diffusion Modelscited by 872
  5. 2021Membership Inference Attacks From First Principlescited by 839
  6. 2024Simple and Effective Masked Diffusion Language Modelscited by 800
  7. 2024Simplified and Generalized Masked Diffusion for Discrete Datacited by 606
  8. 2021Argmax Flows and Multinomial Diffusion: Learning Categorical Distributionscited by 566
  9. 2022DiffuSeq: Sequence to Sequence Text Generation with Diffusion Modelscited by 439
  10. 2022A Continuous Time Framework for Discrete Denoising Modelscited by 437
  11. 2022DeepSpeed Inference: Enabling Efficient Inference of Transformer Models at Unprecedented Scalecited by 403
  12. 2023The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"cited by 399
  13. 2023Detecting Pretraining Data from Large Language Modelscited by 375
  14. 2024Your Absorbing Discrete Diffusion Secretly Models the Conditional Distributions of Clean Datacited by 371
  15. 2022Analog Bits: Generating Discrete Data using Diffusion Models with Self-Conditioningcited by 362
  16. 2024Discrete Flow Matchingcited by 301
  17. 2024Scaling Diffusion Language Models via Adaptation from Autoregressive Modelscited by 268
  18. 2023Membership Inference Attacks against Language Models via Neighbourhood Comparisoncited by 249
  19. 2023Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theorycited by 218
  20. 2022DiffusionBERT: Improving Generative Masked Language Models with Diffusion Modelscited by 214
  21. 2021Autoregressive Diffusion Modelscited by 208
  22. 2024Masked Diffusion Models are Secretly Time-Agnostic Masked Models and Exploit Inaccurate Categorical Samplingcited by 204
  23. 2022Continuous diffusion for categorical datacited by 202
  24. 2024Scaling up Masked Diffusion Models on Textcited by 195
  25. 2022SSD-LM: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Controlcited by 191
  26. 2024Do Membership Inference Attacks Work on Large Language Models?cited by 188
  27. 2022Quantifying Privacy Risks of Masked Language Models Using Membership Inference Attackscited by 168
  28. 2022Score-based Continuous-time Discrete Diffusion Modelscited by 167
  29. 2023A Reparameterized Discrete Diffusion Model for Text Generationcited by 152
  30. 2023Likelihood-Based Diffusion Language Modelscited by 147
  31. 2022Concrete Score Matching: Generalized Score Matching for Discrete Datacited by 144
  32. 2021On the Importance of Difficulty Calibration in Membership Inference Attackscited by 137
  33. 2022Self-conditioned Embedding Diffusion for Text Generationcited by 114
  34. 2023AR-Diffusion: Auto-Regressive Diffusion Model for Text Generationcited by 113
  35. 2024Min-K%++: Improved Baseline for Detecting Pre-Training Data from Large Language Modelscited by 102
  36. 2022Membership Inference Attacks by Exploiting Loss Trajectorycited by 93
  37. 2023Practical Membership Inference Attacks against Fine-tuned Large Language Models via Self-prompt Calibrationcited by 90
  38. 2023Reflected Diffusion Modelscited by 81

50 most cited of 79 All 79 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 5.

5 papers.

  1. 2026Probing Memorization of Tabular In-Context Learningcited by 1
  2. 2026Membership Inference Attacks on Discrete Diffusion Language Modelscited by 0
  3. 2026Backdooring Masked Diffusion Language Modelscited by 0
  4. 2026Extracting Training Data from Diffusion Language Models via Infillingcited by 0
  5. 2026JUMP: Single-Pass Membership Inference on Fine-Tuned Diffusion Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Membership Inference in Fine-tuned Diffusion Language Models via Token-level Memorization Asymmetry
  2. 2025Toward Safer Diffusion Language Models: Discovery and Mitigation of Priming Vulnerability
  3. 2026BadDLM: Backdooring Diffusion Language Models with Diverse Targets
  4. 2025Membership Inference Attacks on Large-Scale Models: A Survey
  5. 2026Noise as a Probe: Membership Inference Attacks on Diffusion Models Leveraging Initial Noise
  6. 2026Adaptive Diffusion Freezing: Privacy-preserving Diffusion Models Against Membership Inference Attacks
  7. 2023Privacy Threats in Stable Diffusion Models
  8. 2026TrustLDM: Benchmarking Trustworthiness in Language Diffusion Models