privacy leakage
- Attractive Metadata Attack: Inducing LLM Agents to Invoke Malicious Tools
- Contextual Integrity in LLMs via Reasoning and Reinforcement Learning
- Deep Learning with Plausible Deniability
- DictPFL: Efficient and Private Federated Learning on Encrypted Gradients
- Differentially Private Federated Low Rank Adaptation Beyond Fixed-Matrix
- Mitigating the Privacy–Utility Trade-off in Decentralized Federated Learning via f-Differential Privacy
- Unlearned but Not Forgotten: Data Extraction after Exact Unlearning in LLM