SzilviaB/SzilviaB_SzilviaB_Solar_Rain-2x10.7b

Full fine-tune vs macadeliccc/SOLAR-2x10.7b-truthy-dpo. Fine-tune: 95% of the weights changed, across 46 of 48 layers.

Tensors changed: lm_head 0/1, Embeddings 0/1, MLP 276/288, Norms 74/97, other 46/48, Attention 184/192

1 derivatives checked

• mradermacher/Solar_Rain-i1-GGUF: Quantization

On the Hub · Downloads on Model Pulse