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

AG

Benchmarks

Task NameDataset NameSOTA ResultTrend
Textual Adversarial AttackAG
Attack Success Rate (ASR)7.5
27
Text ClassificationAG
Accuracy95.5
23
Predicate ClassificationAG dataset
mR@1061
11
Text ClassificationAG (test)
Test Error Rate0.0445
9
Sentiment AnalysisAG (test)
Accuracy92.5
8
Text ClassificationAG (test)
Params (M)2.46
7
Text ClassificationAG (train)
Training Time (h)1
6
Faithfulness DiagnosticityAG (test)
Diagnosticity64.9
4
Showing 8 of 8 rows