kykim0/OLMo-1B-SFT-hf
Full fine-tune vs allenai/OLMo-1B-hf. Fine-tune: 100% of the weights changed, across 16 of 16 layers.
Tensors changed: Embeddings 1/1, MLP 48/48, Attention 64/64
1 derivatives checked
- mradermacher/OLMo-1B-SFT-hf-GGUF: Same weights, converted