win10/DeepSeek-R1-Distill-sthenno-14b-0121-union-tokenizer
Full fine-tune vs sthenno-com/miscii-14b-1225. Fine-tune: 100% of the weights changed, across 48 of 48 layers.
Tensors changed: Norms 97/97, MLP 144/144, Attention 336/336
Full fine-tune vs sthenno-com/miscii-14b-1225. Fine-tune: 100% of the weights changed, across 48 of 48 layers.
Tensors changed: Norms 97/97, MLP 144/144, Attention 336/336