Tag
SHAPE proposes a coalition-aware expert pruning framework for sparse MoE LLMs that uses Shapley-style attribution over routing traces to identify essential experts, achieving competitive accuracy under 20-40% pruning and reducing GPU memory footprint.
This paper introduces Quantum Frog, a two-player cooperative game with a quantized-time mechanic, and uses reinforcement learning to analyze difficulty scaling, optimal strategies, and emergent cooperation between agents.