Statistical Estimation of Adversarial Risk in Large Language Models under Best-of-N Sampling

AAMLRe-assign community