Adversarial Robustness on AdvGLUE

Metric: Accuracy (higher is better)

LeaderboardDataset

Loading chart...

Results

Submit a result

#	Model↕	Accuracy▼	Extra Data	Paper	Date↕	Code
1	DeBERTa (single model)	0.6086	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
2	ALBERT (single model)	0.5922	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
3	T5 (single model)	0.5682	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
4	SMART_RoBERTa (single model)	0.5371	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
5	FreeLB (single model)	0.5048	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
6	RoBERTa (single model)	0.5021	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
7	InfoBERT (single model)	0.4603	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
8	ELECTRA (single model)	0.4169	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
9	BERT (single model)	0.3369	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code
10	SMART_BERT (single model)	0.3029	No	Adversarial GLUE: A Multi-Task Benchmark for Rob...	2021-11-04	Code