Ting Liu

Diagnosing and Remedying Knowledge Deficiencies in LLMs via Label-free Curricular Meaningful Learning

AutoTool: Automatic Scaling of Tool-Use Capabilities in RL via Decoupled Entropy Constraints

CAI: Caption-Sensitive Attention Intervention for Mitigating Object Hallucination in Large Vision-Language Models

Persuade with Reason: Enhancing Debate Persuasiveness through Accurate Persuasion Feedback Derived from Weak Supervised Labels

When Less Language is More: Language-Reasoning Disentanglement Makes LLMs Better Multilingual Reasoners

How Does Sequence Modeling Architecture Influence Base Capabilities of Pre-trained Language Models? Exploring Key Architecture Design Principles to Avoid Base Capabilities Degradation

Teaching Language Models to Evolve with Users: Dynamic Profile Modeling for Personalized Alignment

Lens: Rethinking Multilingual Enhancement for Large Language Models

UFO-RL: Uncertainty-Focused Optimization for Efficient Reinforcement Learning Data Selection

Towards Comprehensive and Efficient Post Safety Alignment of Large Language Models via Safety Patching

Towards Secure Tuning: Mitigating Security Risks Arising from Benign Instruction Fine-Tuning