Use of the original Llama was restricted to researchers. Llama 2 is a pre-trained LLM which has also been refined using reinforcement learning with human feedback (RLHF). The training data ...