Bootstrap knowledge of LLMs ASAP. With a bias/focus to GPT.
Avoid being a link dump. Try to provide only valuable well tuned information.
Neural network links before starting with transformers.
[LLM Training and Fundamentals] | |
- GPT and GPT-2: https://cameronrwolfe.substack.com/p/language-models-gpt-and-gpt-2 | |
- GPT-3 and LLM Scaling: https://cameronrwolfe.substack.com/p/language-model-scaling-laws-and-gpt | |
- Modern LLMs: https://cameronrwolfe.substack.com/p/modern-llms-mt-nlg-chinchilla-gopher | |
- Specialized LLMs: https://cameronrwolfe.substack.com/p/specialized-llms-chatgpt-lamda-galactica | |
[Open Source LLMs] | |
- LLaMA: https://cameronrwolfe.substack.com/p/llama-llms-for-everyone | |
- Beyond LLaMA (Imitation Models): https://cameronrwolfe.substack.com/p/beyond-llama-the-power-of-open-llms | |
- False Promise of Imitation: https://cameronrwolfe.substack.com/p/imitation-models-and-the-open-source |