view article Article Token Merging for fast LLM inference : Background and first trials with Mistral By samchain • Apr 30 • 3