Zeming Wei

Revisiting Hallucination Detection Through The Lens Of Effective Rank-based Uncertainty

Monitoring LLM-based Multi-Agent Systems Against Corruption Attacks via Node Evaluation

Unveiling the Basin-Like Loss Landscape in Large Language Models

Mitigating Fine-tuning Risks in LLMs via Safety-Aware Probing Optimization

Autoregressive Models Rival Diffusion Models at ANY-ORDER Generation

Dynamic Orthogonal Continual Fine-tuning for Mitigating Catastrophic Forgetting of LLMs

Decoding Large Language Diffusion Models with Foreseeing Movement

Towards the Worst-case Robustness of Large Language Models

Advancing LLM Safe Alignment with Safety Representation Ranking

Identifying and Understanding Cross-Class Features in Adversarial Training