lunahr/gemma-3-1b-it-abliterated
Abliterated vs google/gemma-3-1b-it. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 23 of 26 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Norms 0/157, MLP 23/78, Attention 23/104
2 derivatives checked
- Luisz21/gemma-3-1b-it-abliterated-Q8_0-GGUF: Quantization
- Luisz21/gemma-3-1b-it-abliterated-Q4_K_M-GGUF: Quantization