llmfan46/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic
Abliterated vs google/gemma-4-26B-A4B-it-qat-q4_0-unquantized. Looks like abliteration (refusal removal): only the output projections (o_proj) changed, in 12 of 30 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/4, MLP 0/231, Norms 0/433, other 0/122, Attention 12/223
10 derivatives checked
- EZForever/gemma-4-26B-A4B-it-qat-uncensored-heretic-UDmerge-GGUF: Quantization
- llmfan46/gemma-4-26B-A4B-it-qat-q4_0-uncensored-heretic-GGUF: Quantization
- mradermacher/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-i1-GGUF: Quantization
- llmfan46/gemma-4-26B-A4B-it-qat-q4_0-uncensored-heretic-NVFP4: Quantization
- ailexleon/gemma-4-26B-A4B-it-qat-uncensored-heretic-mlx-vlm-4Bit: Quantization
- llmfan46/gemma-4-26B-A4B-it-qat-q4_0-uncensored-heretic-NVFP4-GGUF: Quantization
- mradermacher/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-GGUF: Quantization
- ailexleon/gemma-4-26B-A4B-it-qat-uncensored-heretic-mlx-lm-4Bit: Quantization
- llmfan46/gemma-4-26B-A4B-it-qat-q4_0-uncensored-heretic-GPTQ-Int4: Quantization
- wazimondo/gemma-4-26B-A4B-it-qat-q4_0-unquantized-uncensored-heretic-GGUF: Quantization