17 papers
Patronus: Identifying and Mitigating Transferable Backdoors in Pre-trained Language Models
Tianhang Zhao, Haodong Zhao, Wei Du +5
The ``Pre-train, then fine-tune'' paradigm has revolutionized Natural Language Processing (NLP). In this context, transferable backdoors pose a severe threat to the Pre-trained Lan…
GradSentry: Gradient Spectral Entropy for Backdoor Sample Filtering in Large Language Model Fine-Tuning
Haodong Zhao, Tianyi Xu, Tianhang Zhao +2
Fine-tuning Large Language Models with untrusted data exposes models to backdoor attacks, where poisoned samples cause targeted misbehavior. Existing sample-filtering defenses rely…
LLM DNA: Tracing Model Evolution via Functional Representations
Zhaomin Wu, Haodong Zhao, Ziyang Wang +3
The explosive growth of large language models (LLMs) has created a vast but opaque landscape: millions of models exist, yet their evolutionary relationships through fine-tuning, di…
EmbTracker: Traceable Black-box Watermarking for Federated Language Models
Haodong Zhao, Jinming Hu, Yijie Bai +6
Federated Language Model (FedLM) allows a collaborative learning without sharing raw data, yet it introduces a critical vulnerability, as every untrustworthy client may leak the re…
GhostEI-Bench: Do Mobile Agents Resilience to Environmental Injection in Dynamic On-Device Environments?
Chiyu Chen, Xinhao Song, Yunkai Chai +7
Vision-Language Models (VLMs) are increasingly deployed as autonomous agents to navigate mobile graphical user interfaces (GUIs). Operating in dynamic on-device ecosystems, which i…
ProtegoFed: Backdoor-Free Federated Instruction Tuning with Interspersed Poisoned Data
Haodong Zhao, Jinming Hu, Zhaomin Wu +7
Federated Instruction Tuning (FIT) enables collaborative instruction tuning of large language models across multiple organizations (clients) in a cross-silo setting without requiri…