Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

global

Benchmarks

Task NameDataset NameSOTA ResultTrend
Correctness PredictionGlobal Pooled Datasets
WP-AUC0.723
12
Signed Distance Function RepresentationGlobal Average
IoU81.7
10
Code GenerationGlobal (test)
Pass@113.33
6
Skill routingglobal
Reward89.45
4
Anomaly DetectionGlobal High-Noise Regime η = 0.9 α = 0.9 (maximum feature corruption)
ROC AUC99
2
Anomaly DetectionGlobal High-Purity Regime (low-noise) η = 0.1 α = 0.1
ROC AUC98.3
2
Anomaly DetectionGlobal Full Experimental Grid (All tested scenarios)
ROC AUC98.7
2
Showing 7 of 7 rows