circle
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
CIRCLE
41.07SIM
60
Cropped OOD circle (test)
97.04SSIM
21
Circle Synthetic (full stream)
0.76G-mean
11
Circle Synthetic
100Specificity
11
Circle
92.8Recall
11
Circle
82.3Average PAUC
11
Circle n=16
100Success Rate
6
Circle n=8
100Success Rate
6
Circle n=4
100Success Rate
6
Circle online downstream setting
1.28Normalized Reward
6
Circle (offline)
1.27Normalized Reward
6
Circle 1-loop
0.22F2FC
3
Circle
90Empirical Coverage
2
Circle synthetic (test)
91Coverage
2
CIRCLE Goal 3
86.27Success Rate (SR)
2
CIRCLE Goal 2
86.69SR
2