Show HN: LLM post-training to speak like GenZ, costing less than a cup of coffee

Category: ai-ml

Tags: llm, fine-tuning, genz-slang

Score: 4.3/10 (Innovation: 4, Technical: 5, Documentation: 5, Utility: 3)

This project fine-tunes a small language model to output GenZ slang using supervised fine-tuning and reinforcement learning, all in Google Colab for under $2. It's a fun, low-cost experiment that explores whether SFT alone suffices or RL adds value, though its utility is limited since frontier models already do this better.

Target audience: ai researchers, machine learning engineers

Repository: https://github.com/aidarbek/genz-qwen · Jupyter Notebook

View on Hacker News