prithivMLmods/Gaea-Opus-14B-Exp

Full fine-tune vs prithivMLmods/Tucana-Opus-14B-r999. Fine-tune: 98% of the weights changed, across 48 of 48 layers.

Tensors changed: Norms 97/97, MLP 141/144, Attention 329/336

2 derivatives checked