SelfCAD: Protecting Your Efficient Reasoning Capabilities via Self-Cautious Insertion
Preprint 2026
Publications
You can also find my articles on my Google Scholar profile.
2026
TrustLDM: Benchmarking Trustworthiness in Language Diffusion Model
ICLR 2026 Trustworthy Workshop (First benchmark for evaluating trustworthiness of language diffusion models)
Decoding Large Language Diffusion Models with Foreseeing Movement
ICLR 2026 DeLTa Workshop
Jailbreak and Guard Aligned Language Models with Only Few In-Context Demonstrations
TPAMI 2026 (Adopted at scale by Anthropic)
On the Adversarial Transferability of Generalized “Skip Connections”
TPAMI 2026 (Journal extension of SGM, original paper cited 400+ times on Google Scholar)
2025
Generalist++: A Meta-learning Framework for Mitigating Trade-off in Adversarial Training
arXiv 2025
Are Smarter LLMs Safer? Exploring Safety-Reasoning Trade-offs in Prompting and Fine-Tuning
arXiv 2025 (First to reveal the safety–reasoning capability trade-off)
2024
Fight Back Against Jailbreaking via Prompt Adversarial Tuning
NeurIPS 2024
TERD: A Unified Framework for Safeguarding Diffusion Models Against Backdoors
ICML 2024 (First backdoor input detection method for diffusion models)
PID: Prompt-Independent Data Protection Against Latent Diffusion Models
ICML 2024
2022
When Adversarial Training Meets Vision Transformers: Recipes from Training to Architecture
NeurIPS 2022 (Spotlight, Top 5%) (First work to improve adversarial robustness of ViTs)
Improving Generative Adversarial Networks via Adversarial Learning in Latent Space
NeurIPS 2022 (Spotlight, Top 5%)
DICE: Domain-attack Invariant Causal Learning for Improved Data Privacy Protection and Adversarial Robustness
SIGKDD 2022
Multi-Task Learning Improves Synthetic Speech Detection
ICASSP 2022