Stuck in the Matrix: Probing Spatial Reasoning in Large Language Models | Dark Hacker News