1 paper
Iris Xu, Guangtao Zeng, Zexue He +5
Large language models (LLMs) have shown strong reasoning and coding capabilities, yet they struggle to generalize to real-world software engineering (SWE) problems that are long-ho…