NeshVerse/Uncensored_Nanbeige-4.1-3B
Abliterated vs Nanbeige/Nanbeige4.1-3B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 22 of 32 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: lm_head 0/1, MLP 19/96, Attention 22/128, Norms 0/65, Embeddings 0/1
2 derivatives checked
- mradermacher/Uncensored_Nanbeige-4.1-3B-i1-GGUF: Quantization
- mradermacher/Uncensored_Nanbeige-4.1-3B-GGUF: Same weights, converted