Edens-Gate/llama-MagPicaro-Big

Full fine-tune vs DS-Archive/L3.3-70B-Magnum-v4-SE. Fine-tune: 99% of the weights changed, across 80 of 80 layers.

Tensors changed: Embeddings 0/1, Norms 0/161, MLP 240/240, Attention 320/320, lm_head 1/1

2 derivatives checked