---
datasets:
- reedmayhew/claude-3.7-sonnet-reasoning
- reedmayhew/gpt-4.5-100x
base_model:
- unsloth/Qwen3-4B-unsloth-bnb-4bit
---

This is a fine-tuned version of Qwen3 4B using one reasoning and one non-reasoning dataset from closed-source LLMs (made available by reedmayhew, thanks!).

The total size of this training dataset is around 300 rows. This model was fine-tuned for 3000 steps.