Dual-Process Scaffold Reasoning for Enhancing LLM Code Debugging
Nov 11, 2025·
,,,·
1 min read
Po-Chung (Bryan) Hsieh
C.-P. Chen
J.-L. Li
M.-C. Chang
This paper proposes a dual-process scaffold reasoning framework for large language model (LLM) code debugging. It reaches an 88.91% pass rate on DebugBench, outperforming prior prompting methods.
My role: first author. This work was done during my LLM engineer internship at the Inventec Artificial Intelligence Center.
Authors
M.S. Student in Electrical and Computer Engineering
I am an M.S. student in Electrical and Computer Engineering at UC San Diego, with a B.S. in Electrical Engineering from National Taiwan University. I build AI agents and reinforcement learning methods for scientific discovery, including multi-agent systems for biomedical feature discovery and policy optimization for discrete generative models, and I previously worked as an LLM engineer intern at the Inventec Artificial Intelligence Center.