ResearchBenchmarksAlgorithmic Reasoning on BELLMAN-FORD 20 nodesAlgorithmic Reasoning on BELLMAN-FORD 20 nodesFollow0.0025Key ValueNE++0.0010870.0108260.0205650.030304Oct 26, 2021Evaluation resultsMethodMethodLinksKey ValuePredecessor AccuracyPredecessor CostNE++Training Strategy=NATraining Strategy=NA2021.100.00252.8—NE++Training Strategy=Mult...Training Strategy=Multi-task2021.100.00352.3—NETraining Strategy=Mult...Training Strategy=Multi-task2021.100.01543.4—NETraining Strategy=NATraining Strategy=NA2021.100.01825.7—NE++Training Strategy=Tran...Training Strategy=Transfer 2-Proc.2021.100.02236.2—NETraining Strategy=Tran...Training Strategy=Transfer Fine-tune2021.100.03867.2—