adversarial perturbation (1) dataset (1) gender bias (1) intra-memory conflict (1) knowledge conflict (1) language (1) misinformation (1) model instability (1) nlp (1) reddit (1) social data science (1) sodas (1) uncertainty (1) uncertainty quantification (1) xai (1)