llmfan46/Omega-Evolution-27B-v2.2-ultra-uncensored-heretic
Abliterated vs ReadyArt/Omega-Evolution-27B-v2.2. Looks like abliteration (refusal removal): only the output projections (out_proj) changed, in 2 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: lm_head 0/1, Embeddings 0/4, MLP 0/304, Attention 2/556, Norms 0/319
3 derivatives checked
- mradermacher/Omega-Evolution-27B-v2.2-ultra-uncensored-heretic-GGUF: Quantization
- mradermacher/Omega-Evolution-27B-v2.2-ultra-uncensored-heretic-i1-GGUF: Quantization
- llmfan46/Omega-Evolution-27B-v2.2-ultra-uncensored-heretic-GGUF: Same weights, converted