sophosympatheia/Electranova-70B-v1.0

Full fine-tune vs Sao10K/Llama-3.3-70B-Vulpecula-r1. Fine-tune: 100% of the weights changed, across 80 of 80 layers.

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

2 derivatives checked