Herd: Using multiple, smaller LLMs to match the performances of proprietary, large LLMs via an intelligent composer

Re-assign community