Naphula/MN-12B-Mag-Mell-R1-Uncensored
Abliterated vs inflatebot/MN-12B-Mag-Mell-R1. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 40 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/81, MLP 40/120, Attention 40/160
2 derivatives checked
- mradermacher/MN-12B-Mag-Mell-R1-Uncensored-i1-GGUF: Quantization
- mradermacher/MN-12B-Mag-Mell-R1-Uncensored-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.