StepGame
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Stepgame k=10
88.1Accuracy
56
Stepgame k=4
93.8Accuracy
56
Stepgame k=3
89.5Accuracy
56
StepGame (test)
100Accuracy (1-hop)
21
StepGame larger k generalization (test)
28.53Accuracy (k=6)
6
StepGame with distracting noise (test)
85.77k=1 Accuracy
6