ResearchBenchmarksProof length optimization on MILFollow43.55ImprovementImProver2.10612.865523.62534.3845Oct 7, 2024Evaluation ResultsMethodMethodLinksImprovementNonempty ImprovementAccuracyImproved AccuracyImProver2024.1043.5544.7610045.94ImProver2024.1030.5456.5610050GPT-4o2024.106.2518.5837.514.42GPT-4o2024.103.727.3813.510