Statistical Estimation of Adversarial Risk in Large Language Models under Best-of-N Sampling Paper • 2601.22636 • Published 3 days ago • 5