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

Samples

Benchmarks

Task NameDataset NameSOTA ResultTrend
Text generation1024 samples
Perplexity (GPT-2)104.44
30
Image Editing100 evaluation samples (test)
L1 Loss0.0146
6
Adversarial Attack Transferability1,000 samples (held-out)
Flip Rate98.8
5
Clustering500-point samples
Sample ARI0.724
3
Showing 4 of 4 rows