Smaller language models can be based on a billion parameters or less—still pretty large, but much smaller than foundational LLMs like ChatGPT and Bard. To get value from generative AI, the path ...