ResearchBenchmarksProof Optimization (Mixed) on MIL aggregated (test)Follow27.31ImprovementImProver2.5588.98415.4121.836Oct 7, 2024Evaluation ResultsMethodMethodLinksImprovementNonempty ImprovementAccuracyImproved AccuracyImProver2024.1027.3139.2410030.55ImProverModel=ImProverModel=ImProver2024.1027.3139.2410030.55GPT-4o2024.103.5123.914.75.11GPT-4oModel=GPT-4oModel=GPT-4o2024.103.5123.914.75.11