Text this: Quality-Aware Exploration Budget Allocation for Cooperative Multi-Agent Reinforcement Learning