lemuralabs/Qwen3.6-27B-Fable-Uncensored
Abliterated vs lemuralabs/Qwen3.6-27B-Fable-Agentic. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 40 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, Norms 0/319, Attention 40/556, MLP 40/304