Naphula/Ancient-Awakening-12B-MPOA

Full fine-tune vs ArliAI/Mistral-Nemo-12B-ArliAI-RPMax-v1.2. Fine-tune: 100% of the weights changed, across 40 of 40 layers.

Tensors changed: Norms 66/81, MLP 120/120, Attention 160/160

1 derivatives checked