1 paper
Sirui Hong, Zhijie Liu, Tengfei Li +3
Evaluating LLM-generated interactive software requires execution in addition to static analysis. The key difficulty is that correctness is a graph-level reachable property over lat…