Nexesenex/Llama_3.x_70b_Electra-Legion_fusion_v2
Full fine-tune vs Steelskull/L3.3-Electra-R1-70b. Fine-tune: 90% of the weights changed, across 80 of 80 layers.
Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 160/161, MLP 220/240, Attention 276/320