Nov 25, 2024 Author(s) Alexandre Marques Manager of Machine Learning Research Mark Kurtz CTO, Neural Magic Dan Alistarh Principal Research Scientist Shubhra Pandit Senior Machine Learning Researcher A Sparse Summary Sparse Foundation Model: The first sparse, highly accurate foundation model built on top of Meta’s Llama 3.1 8B with 98% recovery on Open LLM Leaderboard v1 and full recovery across fine-tuning tasks, including math, coding, and chat. Hardware-Accelerated Sparsity: Features a 2:4 spa...