While current LLMs excel at generating code that works, can we trust that generated code in real-world applications? Often, the answer is NO. Since current research has largely focused on functional correctness, leaving code efficiency as a significant bottleneck.
NeurIPS 2025 Spotlight Co-Evolving LLM Coder and Unit Tester via Reinforcement Learning
📄 This repository contains the implementation for the ICSE 2025 paper, "RLCoder: Reinforcement Learning for Repository-Level Code Completion".
The main features of deepsoftwareanalytics/rlcoder are: Software Engineering.
Open-source alternatives to deepsoftwareanalytics/rlcoder include: elfsong/afterburner — While current LLMs excel at generating code that works, can we trust that generated code in real-world applications?… gen-verse/cure — [NeurIPS 2025 Spotlight] Co-Evolving LLM Coder and Unit Tester via Reinforcement Learning. tomsawyerhu/apr-rl — GRPO for Automated Program Repair.
