elinas/Llama-3-15B-Instruct-zeroed-ft
Abliterated vs elinas/Llama-3-15B-Instruct-zeroed. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 64 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/129, MLP 64/192, Attention 64/256