August 11, 2025

Shaping the Future of Large Language Models

Latest

TopGPT

Aug 10, 2025•18 min read

A practical workflow for adapting GPT-2 to emulate a specific influencer’s tone: continue pretraining on long-form transcripts, align with supervised fine-tuning on crafted Q&A, then polish with Direct Preference Optimization (DPO) to balance authenticity and safety.

Finetuning Llama 3.2 3B Instruct

Jul 2, 2025•9 min read

Supervised finetuning Llama model for better at customer servicing.

Breaking Down the Transformer Architecture

Apr 23, 2025•19 min read

A lot of theory about trasnformer and its inner working.

View all posts →

About this blog

Opinionated writing in its purest form—no clutter, no distractions. Just focused ideas, set in a calm, motonchrome frame.