huihui-ai/Huihui-Qwen3-Coder-Next-abliterated
Abliterated vs Qwen/Qwen3-Coder-Next. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 48 of 48 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/157, Attention 48/264, MLP 837/2577, lm_head 0/1
15 derivatives checked
- bartowski/huihui-ai_Qwen3-Coder-Next-abliterated-GGUF: Quantization
- mradermacher/Huihui-Qwen3-Coder-Next-abliterated-i1-GGUF: Quantization
- mradermacher/Huihui-Qwen3-Coder-Next-abliterated-GGUF: Quantization
- Laerins/Huihui-Qwen3-Coder-Next-abliterated-i1-GGUF: Quantization
- Riyan200324200324/Qwen3-Coder-Next-abliterated-GGUF: Quantization
- Rixf123/Lancer-Coder-Next-abliterated-GGUF: Quantization
- Slaxxis/huihui-ai_Qwen3-Coder-Next-ablisdterated-GGUF: Quantization
- nightmedia/Huihui-Qwen3-Coder-Next-abliterated-mxfp8-mlx: Quantization
- veeceey/Huihui-Qwen3-Coder-Next-abliterated-mlx-4bit: Quantization
- McG-221/Qwen3-Coder-Next-abliterated-mxfp4-gs32: Quantization
- magicshim/Huihui-Qwen3-Coder-Next-abliterated-int4-AutoRound: Quantization
- AITRADER/Huihui-Qwen3-Coder-Next-abliterated-mlx-8Bit: Quantization
- Eldadalbajob/Huihui-Qwen3-Coder-Next-abliterated-mlx-4Bit: Quantization
- JaQuan2k6/Huihui-Qwen3-Coder-Next-abliterated-4bit-mlx: Quantization
- AITRADER/Huihui-Qwen3-Coder-Next-abliterated-mlx-4Bit: Quantization
1 other repos hold byte-for-byte identical weight files.