mergekit-community/mergekit-slerp-rxkhjnf
Full fine-tune vs NousResearch/Meta-Llama-3-8B-Instruct. Fine-tune: 94% of the weights changed, across 30 of 32 layers.
Tensors changed: Norms 60/65, MLP 90/96, Attention 120/128
1 derivatives checked
- MaziyarPanahi/mergekit-slerp-rxkhjnf-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.