Hi there, I'm Dong Jun Kim 👋
I'm currently pursuing my Ph.D. at at Korea University's NLP&AI Lab, specializing in Mechanistic Interpretability of Large Language Models (LLMs). My research focuses on understanding the internal workings of LLMs, developing methods to make their decision-making processes more transparent and interpretable. I am passionate about advancing the field of AI by improving our ability to analyze and explain model behavior at a granular level.
Research Interests:
- Mechanistic Interpretability of LLMs
- Sparse Autoencoders & LLM Circuits
- Retrieval-Augmented Generation (RAG)
- LLM Architectures & Optimization
- Efficient Fine-tuning Techniques
Open to:
- Research collaborations in AI/ML
- Industry partnerships for LLM development
- Reviewer or PC member roles for AI/ML conferences or journals
- Mechanistic Interpretability: Developing novel techniques to understand the internal structures and decision-making pathways in LLMs.
- Sparse Autoencoders & LLM Circuits: Researching sparse representations and circuit-level analysis within LLMs to enhance interpretability.
- LLM Architectures: Designing and optimizing large-scale architectures for efficiency and performance.
- Retrieval-Augmented Generation (RAG): Integrating retrieval mechanisms with generative models to improve accuracy and relevance in language generation tasks.
