3MPER0RR/Ornith1.5-9B-3MPER0RR-abliterated
Abliterated vs ornith-ai/Ornith-1.5-9B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 23 of 32 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/4, Norms 0/215, Attention 23/332, MLP 19/208, lm_head 0/1
4 derivatives checked
- mradermacher/Ornith1.5-9B-3MPER0RR-obliterated-i1-GGUF: Quantization
- mradermacher/Ornith1.5-9B-3MPER0RR-abliterated-i1-GGUF: Quantization
- mradermacher/Ornith1.5-9B-3MPER0RR-obliterated-GGUF: Same weights, converted
- mradermacher/Ornith1.5-9B-3MPER0RR-abliterated-GGUF: Same weights, converted