作者位次 #2/3ACL 2026Findings· Findings of ACL
LLM Program Optimization via Retrieval Augmented Search
Sagnik Anupam · Alexander Shypula · Osbert Bastani
摘要
Recent work has demonstrated the potential of large language models (LLMs) for program optimization, a key challenge in programming languages. We propose a blackbox adaptation method called Retrieval Augmented Search (RAS) that performs beam search over candidate optimizations; at each step, it r...