grimjim/mistralai-Mistral-Nemo-Instruct-2407-extensive-BP-abliteration-12B
Abliterated vs mistralai/Mistral-Nemo-Instruct-2407. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 35 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/81, MLP 35/120, Attention 35/160, lm_head 0/1
4 derivatives checked
- mradermacher/mistralai-Mistral-Nemo-Instruct-2407-extensive-BP-abliteration-12B-i1-GGUF: Quantization
- mradermacher/mistralai-Mistral-Nemo-Instruct-2407-extensive-BP-abliteration-12B-GGUF: Quantization
- xs7david/mistralai-Mistral-Nemo-Instruct-2407-extensive-BP-abliteration-12B-Q4_K_S-GGUF: Quantization
- xs7david/mistralai-Mistral-Nemo-Instruct-2407-extensive-BP-abliteration-12B-Q2_K-GGUF: Quantization