The instruction following is one of the cornerstones of the current generation of large language models(LLMs). Reinforcement learning with human preferences)RLHF) and techniques such as InstructGPT have been the core foundation of breakthroughs such as ChatGPT and GPT-4. However, these models remain hidden behind APIs and we know very little about its underlying architecture. Recently, the from data infrastructure powerhouse Databricks open-sourced Dolly, an instruction-following model that exhibits similar characteristics so ChatGPT despite being much smaller. A few days after, they released Dolly 2.0, the first commercial version of the model. ...