From Building Blocks to Planning: Multi-Step Spatial Reasoning in LLMs with Reinforcement Learning

LRMRe-assign community