[Sys] Find More LLM for Experiments

[slark-prime]

• [ ] Explore what model can be used for Question Answering Tasks(Instruct-Fine-tuned, QA specific, etc)
• [ ] Find smaller LLMs with input lengths similar to GPT-3 (2,048 tokens);