richardyoung/qwen3-32b
Abliterated vs Qwen/Qwen3-32B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 57 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/257, MLP 57/192, Attention 45/256, lm_head 0/1