Grouped Value Attention: Efficient KV Caching via On-Demand Key Reconstruction Paper • 2609.13285 • Published 18 days ago • 81
Running on Zero MCP 2 Lumma 0.6B Tool Playground 🛠 2 Define JSON tools and watch Lumma-0.6B generate tool calls
Grouped Query Experts: Mixture-of-Experts on GQA Self-Attention Paper • 2606.20945 • Published Jun 18 • 81