An Analysis for Reasoning Bias of Language Models with Small Initialization.

Junjie Yao, Zhongwang Zhang, Zhi-Qin John Xu

Browse the full ICML paper archive.