Nabbers1999/Qwen2.5-Coder-32B-Instruct-NP-Abliterated
Abliterated vs Qwen/Qwen2.5-Coder-32B-Instruct. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 63 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/129, MLP 63/192, Attention 63/448, lm_head 0/1