Dist2ill: Distributional Distillation for One-Pass Uncertainty Estimation in Large Language ModelsPublished in Preprint, 2026Share on Twitter Facebook LinkedIn Previous Next