Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
TextFoolergithub.com · Low evidenceTextFooler is a model for generating natural language adversarial attacks against text classification and inference systems. It provides source code and pre-trained model parameters to test the robustness of models like BERT, LSTM, and CNN across seven datasets.
Parse Score