huihui-ai/Huihui-Ornith-1.5-9B-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 21 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 21/332, MLP 21/208, lm_head 0/1
9 derivatives checked
- mradermacher/Huihui-Ornith-1.5-9B-abliterated-i1-GGUF: Quantization
- mradermacher/Huihui-Ornith-1.5-9B-abliterated-GGUF: Same weights, converted
- osxest/Huihui-Ornith-1.5-9B-abliterated-mlx-4Bit: Quantization
- sli3/Cyber-Ornith-1.5-9B-GGUF: Quantization
- keXjos/Huihui-Ornith-1.5-9B-abliterated-mlx-5Bit: Quantization
- osxest/Huihui-Ornith-1.5-9B-abliterated-mlx-6Bit: Quantization
- osxest/Huihui-Ornith-1.5-9B-abliterated-mlx-8Bit: Quantization
- shoemoney/Ornith-1.5-9B-Abliterated-MLX-mxfp4: Quantization
- shoemoney/Ornith-1.5-9B-Abliterated-MLX-mixed36: Quantization
2 other repos hold byte-for-byte identical weight files.