Evaluation Tasks
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Evaluation Tasks Zero-shot Aggregate
79.95Avg. Accuracy
74
Evaluation Tasks Zero-shot Average
60.47Zero-shot Average Accuracy
17
10,000 Evaluation Tasks Difficult n=2563
0.32Path Length Mean (m)
3
10,000 Evaluation Tasks Medium n=4795
0.57Mean Path Length (m)
3
10,000 Evaluation Tasks Easy n=2642
0.53Path Length Mean (m)
3
Evaluation Tasks n=10000 (All)
0.45Mean Path Length (m)
3