Goal-Guided Reinforcement Learning: Leveraging Large Language Models for Long-Horizon Task Decomposition
