tostideluxekaas/Llama-3.2-3B-Instruct-uncensored
Abliterated vs meta-llama/Llama-3.2-3B-Instruct. 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 18/84, Attention 22/112
1 derivatives checked
- tostideluxekaas/Llama-3.2-3B-Instruct-uncensored-GGUF: Same weights, converted