ResearchBenchmarksNatural Language Inference on MNLI (test) (Adversarial Robustness)Natural Language Inference on MNLI (test) (Adversarial Robustness)Follow84.4Clean AccuracyFine-tune82.8483.24583.6584.055Nov 6, 2022Evaluation resultsMethodMethodLinksClean AccuracyAua (TextFooler)Aua (TextBugger)Fine-tune2022.1184.47.74.3InfoBERT2022.1184.110.88.4RobustTSparsity=30%Sparsity=30%2022.118418.422.6FreeLB2022.1182.9118.4