ResearchBenchmarksAlgorithmic Reasoning on BELLMAN-FORD 100 nodesAlgorithmic Reasoning on BELLMAN-FORD 100 nodesFollow1,980,000Key ValueNE-79,190.8064455,406.8068990,004.421,524,602.0332Oct 26, 2021Evaluation resultsMethodMethodLinksKey ValuePredecessor AccuracyNETraining Strategy=NATraining Strategy=NA2021.101,980,00026.1NETraining Strategy=Tran...Training Strategy=Transfer Fine-tune2021.10172,00024.2NETraining Strategy=Mult...Training Strategy=Multi-task2021.101,5309.6NE++Training Strategy=NATraining Strategy=NA2021.101969.5NE++Training Strategy=Tran...Training Strategy=Transfer 2-Proc.2021.1010.816.8NE++Training Strategy=Mult...Training Strategy=Multi-task2021.108.8413