An LLM inside an LLM - Same guy who put a RISK-V CPU inside Qwen has now made Karpathy's stories260k run on an NPU compiled inside Qwen 3.5-4B's weights

A developer has successfully implemented an LLM, Karpathy's stories260k, to run on an NPU compiled within the weights of another LLM, Qwen 3.5-4B. This follows their previous work of integrating a RISC-V CPU inside Qwen. The project, named QwenPU, is available on GitHub and Hugging Face, with further optimizations and a full write-up anticipated soon.