Kooten/gemma-4-31b-abliterated-v1.1
Abliterated vs google/gemma-4-31B-it. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 40 of 60 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/4, Norms 0/523, other 0/62, MLP 40/180, Attention 27/230
4 derivatives checked
- mradermacher/gemma-4-31b-abliterated-v1.1-i1-GGUF: Quantization (fewer layers)
- Kooten/gemma-4-31b-abliterated-v1.1-GGUF: Quantization (fewer layers)
- mradermacher/gemma-4-31b-abliterated-v1.1-GGUF: Quantization (fewer layers)
- Enerccio/gemma-4-31b-abliterated-GGUF: Quantization (fewer layers)
1 other repos hold byte-for-byte identical weight files.