nitky/Superswallow-70b-v0.3
Full fine-tune vs allenai/tulu-2-dpo-70b. Fine-tune: 100% of the weights changed, across 80 of 80 layers.
Tensors changed: Norms 161/161, MLP 240/240, Attention 320/320
3 derivatives checked
- mradermacher/Superswallow-70b-v0.3-i1-GGUF: Quantization
- mradermacher/Superswallow-70b-v0.3-GGUF: Quantization
- nitky/Superswallow-70b-RP-v0.3: LoRA-style