prithivMLmods/Lalande-21185-Qwen-Exp

Full fine-tune vs prithivMLmods/Calcium-Opus-14B-Elite2. Fine-tune: 100% of the weights changed, across 48 of 48 layers.

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

1 derivatives checked