---
tags:
- unsloth
---
### Content
Using Rayman-Extraction-Dataset to train smolLM 360M to see if it is capable of performing data extraction on par with Llama 3.2 1B & 3B finetunes