Semantic Membership Inference Attack against Large Language Models

2024/06/14 by Hamid Mozaffari, Mozaffari, Hamid, Virendra J. Marathe +1 · 6 citations
Computer Science · Medicine · #Adversarial Robustness in Machine Learning #Artificial Intelligence in Healthcare and Education #FOS: Computer and information sciences #Machine Learning (cs.LG) #Topic Modeling

paper · pdf · doi:10.48550/arxiv.2406.10218

openalex publication_date 2024/06/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Membership Inference Attacks (MIAs) determine whether a specific data point was included in the training set of a target model. In this paper, we introduce the Semantic Membership Inference Attack (SMIA), a novel approach that enhances MIA performance by leveraging the semantic content of inputs and their perturbations. SMIA trains a neural network to analyze the target model's behavior on perturbed inputs, effectively capturing variations in output probability distributions between members and non-members. We conduct comprehensive evaluations on the Pythia and GPT-Neo model families using the Wikipedia dataset. Our results show that SMIA significantly outperforms existing MIAs; for instance, SMIA achieves an AUC-ROC of 67.39% on Pythia-12B, compared to 58.90% by the second-best attack.

Cited by

• PerProb: Indirectly Evaluating Memorization in Large Language Models
• FiMMIA: scaling semantic perturbation-based membership inference across modalities
• Memories Retrieved from Many Paths: A Multi-Prefix Framework for Robust Detection of Training Data Leakage in Large Language Models
• Multimodal Evaluation of Russian-language Architectures
• EL-MIA: Quantifying Membership Inference Risks of Sensitive Entities in LLMs
• LLM in the Middle: A Systematic Review of Threats and Mitigations to Real-World LLM-based Systems

Related

• Do Membership Inference Attacks Work on Large Language Models?
• Context-Aware Membership Inference Attacks against Pre-trained Large Language Models
• Towards Label-Only Membership Inference Attack against Pre-trained Large Language Models
• Detecting Training Data of Large Language Models via Expectation Maximization
• Scaling Up Membership Inference: When and How Attacks Succeed on Large Language Models