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

Real Data

Benchmarks

Task NameDataset NameSOTA ResultTrend
Bandit LearningReal Data (Attack) Wine Quality
Mean Regret1,620.39
10
Novel View SynthesisReal Data (Interpolation)
PSNR32.653
9
Factual Accuracy EstimationReal-data Factual Mistral responses
Original Accuracy51.05
6
Factual Accuracy EstimationReal-data Factual Qwen responses
Original Accuracy55.94
6
Factual Accuracy EstimationReal-data Factual Gemini responses
Original Accuracy66.37
6
Factual Accuracy EstimationReal-data Factual GPT-5.4-mini responses
Original Accuracy62.47
6
Factual Accuracy EstimationReal-data Factual GPT-5.4 responses
Original Accuracy64.39
6
cube-single-playreal-data-E
Normalized Score32.7
6
cube-single-playreal-data-C electricity-hourly
Normalized Score42.1
6
cube-single-playreal-data-A australian-electricity-demand
Normalized Score23.7
6
Shape PredictionReal Data
Chamfer Distance (mm)3.786
3
Similarity Search (Fine-grained match)Real data
Precision@347.4
3
Similarity Search (Family match)Real data
Precision@364.5
3
Object Retrieval & Goal NavigationReal Data unmapped UO
R-RSR Success Rate1
3
Object Retrieval & Goal NavigationReal Data relocated RO
R-RSR1
3
Object Retrieval & Goal NavigationReal Data static SO
R-RSR1
3
Synthetic Data Quality AssessmentReal Data reference
FID14.3
2
Congruence of heading alignment distributionsReal data (test)
Wasserstein-1 (W1)0.0458
1
Showing 18 of 18 rows