MuXodious/Oneirogen-7B-abominable-heresy

Abliterated vs gustavecortal/Oneirogen-7B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 22 of 28 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.

Tensors changed: Embeddings 0/1, Norms 0/57, MLP 7/84, Attention 22/196, lm_head 0/1

2 derivatives checked

On the Hub · Downloads on Model Pulse