bamec66557/MISCHIEVOUS-12B-Mix_Neo
Partial fine-tune vs bamec66557/MISCHIEVOUS-12B-Mix_0.4v. Partial change: 50% of the weights changed (attention, embeddings, mlp, norms), in 20 of 40 layers.
Tensors changed: lm_head 0/1, Embeddings 1/1, Norms 18/81, MLP 60/120, Attention 80/160
2 derivatives checked
- mradermacher/MISCHIEVOUS-12B-Mix_Neo-i1-GGUF: Quantization
- mradermacher/MISCHIEVOUS-12B-Mix_Neo-GGUF: Quantization