Evaluating Allocational Harms in Large Language Models Code for our paper The Mismeasure of Man and Models: Evaluating Allocational Harms in Large Language Models