Research map / Adversarial robustness and security / Attacks on text and graphs
Membership Inference Attacks Against Fine-tuned Diffusion Language Models
Yuetian Chen, Kaiyuan Zhang, Yuntao Du, Edoardo Stoppa, Charles Fleming, Ashish Kundu, Bruno Ribeiro, Ninghui Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use SAMA sign-based subset aggregation instead of loss/ratio attacks | LLaDA-8B-Base fine-tuned · MIMIR ArXiv | AUC: 0.597 -> 0.850 | Improved |
| use SAMA instead of Ratio | LLaDA-8B-Base fine-tuned · MIMIR GitHub | TPR@10%FPR: 0.355 -> 0.647 | Improved |
| use SAMA instead of Ratio on HackerNews | LLaDA-8B-Base fine-tuned · MIMIR HackerNews | AUC: 0.575 -> 0.657 | Improved |
| apply autoregressive MIAs (Loss, Min-K%) to DLMs | LLaDA-8B-Base fine-tuned · MIMIR | AUC: 0.506 (Loss, ArXiv) | No clear effect |
| use diffusion-image MIAs SecMI/PIA | LLaDA-8B-Base fine-tuned · MIMIR | AUC: 0.520 / 0.525 (ArXiv) | No clear effect |
| add reference model calibration to loss attack (ablation) | SAMA · MIMIR | AUC: +0.09 to +0.19 AUC | Improved |
| add progressive masking (ablation) | SAMA · MIMIR | AUC: 2-3% gains | Improved |
| add robust subset sign-based aggregation (ablation) | SAMA · MIMIR | AUC: 20-30% AUC increase | Improved |
| add inverse-step adaptive weighting (ablation) | SAMA · MIMIR | AUC: 3-5% refinement | Improved |
Closest papers by meaning
- Exploring the limits of strong membership inference attacks on large language models (NeurIPS 2025 · similarity 0.90)
- MaskForge: Structure-Aware Adaptive Attacks for Jailbreaking Diffusion Large Language Models (NeurIPS 2026 · similarity 0.88)
- An Efficient Membership Inference Attack for the Diffusion Model by Proximal Initialization (ICLR 2024 · similarity 0.88)
- Privacy Auditing of Large Language Models (ICLR 2025 · similarity 0.88)
- Membership Inference Attacks against Fine-tuned Large Language Models via Self-prompt Calibration (NeurIPS 2024 · similarity 0.88)
- Re-evaluating Confidence Remasking in Masked Diffusion Language Models (NeurIPS 2026 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
