Sleeping Agent Super LLM (Qwen2.5-1.5B-Instruct, CPU) 🧠Coherent, responsive CPU chat (Qwen2.5-1.5B-Instruct)