This repository documents my journey of learning, experimenting with, and implementing Large Language Models (LLMs).
Notebook Viewer: View interactive code diff notebooks at https://app.reviewnb.com/aki-008/LLM_track/
Quick Links:
- Main notebook viewer above
- Experiment implementations in individual folders
Instead of only using existing frameworks and APIs, I want to understand how modern LLMs work under the hood by building, testing, and improving different components myself.
Instead of only using existing frameworks and APIs, I want to understand how modern LLMs work under the hood by building, testing, and improving different components myself.
The repository includes:
- Implementations of LLM concepts and architectures
- Experiments with training, fine-tuning, and optimization
- Reproductions of techniques from recent research papers
- Notes and explanations from my learning process
- Small projects and benchmarks to evaluate new ideas
My goal is to continuously implement techniques introduced by frontier AI research labs and the open-source community, both to deepen my understanding and to demonstrate my ability to translate research into working code.
This repository serves as a public record of my progress, where each commit reflects something new I've learned or built. As I continue reading papers and experimenting with new ideas, I'll keep updating this repository with implementations, improvements, and observations.
Learning in public, one implementation at a time.