Picture for Xuebing Zhou

Xuebing Zhou

PrivacyScalpel: Enhancing LLM Privacy via Interpretable Feature Intervention with Sparse Autoencoders

Add code
Mar 14, 2025
Viaarxiv icon

PII-Scope: A Benchmark for Training Data PII Leakage Assessment in LLMs

Add code
Oct 09, 2024
Figure 1 for PII-Scope: A Benchmark for Training Data PII Leakage Assessment in LLMs
Figure 2 for PII-Scope: A Benchmark for Training Data PII Leakage Assessment in LLMs
Figure 3 for PII-Scope: A Benchmark for Training Data PII Leakage Assessment in LLMs
Figure 4 for PII-Scope: A Benchmark for Training Data PII Leakage Assessment in LLMs
Viaarxiv icon

Privacy Checklist: Privacy Violation Detection Grounding on Contextual Integrity Theory

Add code
Aug 19, 2024
Figure 1 for Privacy Checklist: Privacy Violation Detection Grounding on Contextual Integrity Theory
Figure 2 for Privacy Checklist: Privacy Violation Detection Grounding on Contextual Integrity Theory
Figure 3 for Privacy Checklist: Privacy Violation Detection Grounding on Contextual Integrity Theory
Figure 4 for Privacy Checklist: Privacy Violation Detection Grounding on Contextual Integrity Theory
Viaarxiv icon

ObfuscaTune: Obfuscated Offsite Fine-tuning and Inference of Proprietary LLMs on Private Datasets

Add code
Jul 03, 2024
Viaarxiv icon

PII-Compass: Guiding LLM training data extraction prompts towards the target PII via grounding

Add code
Jul 03, 2024
Viaarxiv icon

IncogniText: Privacy-enhancing Conditional Text Anonymization via LLM-based Private Attribute Randomization

Add code
Jul 03, 2024
Viaarxiv icon