Text this: Rethinking Exploration and Experience Exploitation in Value-Based Multi-Agent Reinforcement Learning