Diving into Prompt Engineering for LLMs within this repository. Learning and practicing the latest techniques and methods through practical examples to gain working knowledge in this ever-growing technology.
Mixtral8x7B-Instruct LLM model is used explore the advanced prompting techniques. This LLM is capable enough to run on google colab thanks to Denis Mazur and Artyom Eliseev who have made this possible by quantizing the original model in mixed precision and implementing a MoE-specific offloading strategy.