saidutta69/Yi-1.5-9B-Chat-16K-heretic
Abliterated vs 01-ai/Yi-1.5-9B-Chat-16K. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 42 of 48 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/97, MLP 39/144, Attention 42/192, lm_head 0/1