PhD student, Peking University
2 papers at NeurIPS 2025
We introduce rStar-Coder to train advanced code reasoning LLMs, with our 14B model achieving comparable performance to QWQ-32B.