wangzhang/gemma-4-26B-A4B-it-abliterix
Partial fine-tune vs google/gemma-4-26B-A4B-it. Partial change: 22% of the weights changed (attention, mlp, other), in 23 of 30 layers.
Tensors changed: Embeddings 0/4, MLP 42/231, Norms 0/433, other 1/122, Attention 23/223
12 derivatives checked
- mradermacher/gemma-4-26B-A4B-it-abliterix-i1-GGUF: Quantization
- mradermacher/gemma-4-26B-A4B-it-abliterix-GGUF: Quantization
- wangzhang/gemma-4-26B-A4B-it-abliterix-GGUF: Abliterated
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-2bit-int2-affine: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-4bit-nvfp4: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-3bit-int3-affine: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-5bit-int5-affine: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-6bit-int6-affine: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-4bit-mixed_4_6: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-3bit-mixed_3_6: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-8bit-int8-affine: Quantization
- zecanard/gemma-4-26B-A4B-it-uncensored-abliterix-MLX-2bit-mixed_2_6: Quantization