mergekit-community/MN-Sappho-jlcj-12B
Partial fine-tune vs mergekit-community/MN-Sappho-c-12B. Partial change: 87% of the weights changed (attention, embeddings, lm_head, mlp, norms), in 34 of 40 layers.
Tensors changed: lm_head 1/1, Embeddings 1/1, Norms 28/81, MLP 102/120, Attention 136/160
1 derivatives checked
- mradermacher/MN-Sappho-jlcj-12B-GGUF: Quantization