darrellbest/Qwen-Image-2.1-PE-T2I-Heretic
Abliterated vs Qwen/Qwen-Image-2.1-PE-T2I. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 27 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 27/332, MLP 21/208, lm_head 0/1
1 derivatives checked
- darrellbest/Qwen-Image-2.1-PE-T2I-Heretic-GGUF: Same weights, converted