ResearchBenchmarksNatural Language to Signal Temporal Logic Generation on DeepSTL Human Evaluation 1.0Follow54.6AccuracyKGST41.49644.89848.351.702May 27, 2025Evaluation ResultsMethodMethodLinksAccuracyKGST2025.0554.6DeepSeek2025.0549.2GPT-42025.0548.8GPT-4+Self-Refinestrategy=Self-Refinestrategy=Self-Refine2025.0547GPT-3.52025.0545.6DeepSTL2025.0542