huihui-ai/Huihui-Qwen3.5-35B-A3B-abliterated
Abliterated vs Qwen/Qwen3.5-35B-A3B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 40 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: MLP 337/1165, lm_head 0/1, Embeddings 0/5, Attention 41/392, other 0/1, Norms 0/247
11 derivatives checked
- mradermacher/Huihui-Qwen3.5-35B-A3B-abliterated-GGUF: Quantization
- mradermacher/Huihui-Qwen3.5-35B-A3B-abliterated-i1-GGUF: Quantization
- HeYujie/Qwen3.5-35B-A3B-abliterated-GGUF: Quantization
- noctrex/Huihui-Qwen3.5-35B-A3B-abliterated-MXFP4_MOE-GGUF: Same weights, converted
- jduartedj/MiniCPM-V-4.6-35B-Abliterated: Same weights, repacked
- cesarsal1nas/Huihui-Qwen3.5-35B-A3B-abliterated-Q4_K_M-GGUF: Quantization
- YTan2000/Huihui-Qwen35-A35B-Ablit-TQ3_4S: Quantization
- mlx-community/Huihui-Qwen3.5-35B-A3B-abliterated-mxfp4: Quantization
- mlx-community/Huihui-Qwen3.5-35B-A3B-abliterated-6bit: Quantization
- cs2764/Huihui-Qwen3.5-35B-A3B-abliterated-Q6_K-GGUF: Quantization
- YTan2000/Huihui-Qwen35-A35B-Ablit-Small-TQ3_4S: Quantization
27 other repos hold byte-for-byte identical weight files.