ResearchBenchmarksProof Optimization (Length) on MIL aggregated (test)Follow20.96ImprovementImProver3.00967.669812.3316.9902Oct 7, 2024Evaluation ResultsMethodMethodLinksImprovementNonempty ImprovementAccuracyImproved AccuracyImProver2024.1020.9655.2910035.44ImProverModel=ImProverModel=ImProver2024.1020.9655.2910035.44GPT-4o2024.103.715.1526.368.31GPT-4oModel=GPT-4oModel=GPT-4o2024.103.715.1526.368.31