Nabbers1999/Mini-Llama-8B-Chat-SFT-0129

Full fine-tune (fewer layers) vs mistralai/Ministral-3-8B-Base-2512. Nabbers1999/Mini-Llama-8B-Chat-SFT-0129 matches part of mistralai/Ministral-3-8B-Base-2512 (307 of its 531 tensors; the rest of mistralai/Ministral-3-8B-Base-2512 has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 34 of 34 layers.

Tensors changed: Norms 1/69, MLP 102/102, Attention 136/136

1 derivatives checked

On the Hub · Downloads on Model Pulse