[short] Can Large Language Models Understand Context?

This paper introduces a context understanding benchmark for evaluating Large Language Models' ability to understand contextual features, finding that pre-trained dense models struggle with nuanced context understanding.