Large Language Models (LLMs) are entirely controllable through human prompts and lack 'emergent abilities'; that is, the means to form their own insights or conclusions. Increasing model size does not ...