1 paper · 1 filter
Chenkai Ma, Xinya Du
Language models (LMs) are capable of conducting in-context learning for multiple choice reasoning tasks, but the options in these tasks are treated equally. As humans often first e…