migtissera/Tess-v2.5-Phi-3-medium-128k-14B

Full fine-tune vs microsoft/Phi-3-medium-128k-instruct. Fine-tune: 100% of the weights changed, across 40 of 40 layers.

Tensors changed: Embeddings 1/1, Norms 81/81, MLP 80/80, Attention 80/80, lm_head 1/1

2 derivatives checked