internetov1488/qwen3-4b-abliterated
Abliterated vs Qwen/Qwen3-4B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 22 of 36 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/145, MLP 18/108, Attention 22/144