Avesed/Qwopus3.6-27B-v2-abliterated
Abliterated vs Jackrong/Qwopus3.6-27B-v2. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 64 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/5, MLP 65/307, Attention 65/560, lm_head 0/1, Norms 0/325, other 0/1
3 derivatives checked
- mradermacher/Qwopus3.6-27B-v2-abliterated-i1-GGUF: Quantization
- mradermacher/Qwopus3.6-27B-v2-abliterated-GGUF: Quantization
- Avesed/Qwopus3.6-27B-v2-abliterated-GGUF: Quantization