Weixiang Zhao

AlphaSteer: Learning Refusal Steering with Principled Null-Space Constraint

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

L-MTP: Leap Multi-Token Prediction Beyond Adjacent Context for Large Language Models

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

On Reasoning Strength Planning in Large Reasoning Models

Lens: Rethinking Multilingual Enhancement for Large Language Models

RSafe: Incentivizing proactive reasoning to build robust and adaptive LLM safeguards

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