Accumulating Context Changes the Beliefs of Language Models

Geng, Jiayi, Chen, Howard, Liu, Ryan · arXiv · 2025