Rongxiang Weng

Expanding Reasoning Potential in Foundation Model by Learning Diverse Chains of Thought Patterns

Two Minds Better Than One: Collaborative Reward Modeling for LLM Alignment

Libra: Assessing and Improving Reward Model by Learning to Think

Make Mathematical Reasoning Adaptive

OneCAT: Decoder-Only Auto-Regressive Model for Unified Understanding and Generation

XBOUND: Exploring Capability Boundaries of Device-Control Agents at the State Level

Length Desensitization in Direct Preference Optimization