lm-provers/QED-Nano-SFT
Full fine-tune vs Qwen/Qwen3-4B-Thinking-2507. Fine-tune: 100% of the weights changed, across 36 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 129/145, MLP 108/108, Attention 144/144
2 derivatives checked
- Mungert/QED-Nano-GGUF: Full fine-tune
- lm-provers/QED-Nano: Full fine-tune