Zynerji/Ektome-SmolLM3-3B-PristinelyUncensored
Abliterated vs HuggingFaceTB/SmolLM3-3B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 36 of 36 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/73, MLP 36/108, Attention 36/144
2 derivatives checked
- mradermacher/Ektome-SmolLM3-3B-PristinelyUncensored-i1-GGUF: Quantization
- mradermacher/Ektome-SmolLM3-3B-PristinelyUncensored-GGUF: Same weights, converted