Igniting Creative Writing in Small Language Models: LLM-as-a-Judge versus Multi-Agent Refined Rewards Paper • 2508.21476 • Published Aug 29, 2025 • 4