OOD
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
OOD Average
63.57Accuracy
22
Four OOD (test)
1.57Macro-F1 Delta
21
OOD 6 tasks (test)
83.7Accuracy
20
OOD Mean
5.21Speedup
20
OOD datasets
94.2pAUROC@20
17
OOD 1.0 (test)
7,824FG-ARI
16
OOD n=176 (test)
75.89Average Performance (%)
14
OOD
89Accuracy
11
OOD n = 26 (test)
93.7DSC
10
OOD Average across 8 domains
68.73AUROC
10
OOD
0.822AUC (Confidence)
9
OOD AIME, HMMT, GPQA, MMLU-Pro, MMLU-Redux 2.0
89.5Pass@1
8
OOD
1.285Loss
7
OOD datasets (test)
55.6Average Target Accuracy
6
OOD
40Overall Match Accuracy
6
Avg. OOD Aggregate of TSDiff, Diffusion-TS, WaveStitch (summary)
84.8F1 Score
6
OOD
54TP Score
6
OOD
65.71Accuracy
6
OOD
2.13Block Efficiency
5
OOD Shopping n = 105 (test)
48Precision
5
OOD set
0.54NMI
5
OOD set
92.7Accuracy
5
OOD out-of-domain (150 samples)
102Count
4
OOD Case II: high obstacle density (30 obstacles)
44.2SR
4
OOD Average
39.94PSNR
4