huihui-ai/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated
Abliterated vs lordx64/Qwen3.6-35B-A3B-Claude-4.7-Opus-Reasoning-Distilled. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 39 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/5, Attention 39/392, MLP 78/399, Norms 0/247, lm_head 0/1, other 0/1
31 derivatives checked
- deadbydawn101/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-GGUF: Full fine-tune
- sakamakismile/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-NVFP4: Quantization
- mradermacher/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-i1-GGUF: Quantization
- deadbydawn101/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-mlx: Partial fine-tune
- whoya/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-Q4_K_M-GGUF: Quantization
- mradermacher/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-GGUF: Quantization
- deadbydawn101/RavenX-CyberAgent-v6.2-Experimental-GGUF: Quantization
- endystrike/Endy-Qwen3.6-CyberSec-35B-A3B-GGUF: Full fine-tune
- mlx-community/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-mlx-4bit-mtp-msq: Quantization
- shennguyen/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-GGUF: Same weights, converted
- ArkhAngelLifeJiggy/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-mlx: Partial fine-tune
- Achilles1089/fable-coder-35B-A3B: Full fine-tune
- Neurona/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-GGUF: Full fine-tune
- IHateCrickets/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-MLX-6bit: Quantization
- ScottzillaSystems/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-GGUF: Full fine-tune
- mlx-community/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-mlx-4bit-msq: Quantization
- whoya/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-Q5_K_M-GGUF: Quantization
- googlecs/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-Q4_K_M-GGUF: Quantization
- Tree3media/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-mlx: Partial fine-tune
- Kiliver/RavenX-CyberAgent-Qwen3.6-35B-A3B-Opus-4.7-OpenMythos-Pentester-BugHunter-RATH-GGUF: Full fine-tune
- vectoronfire/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-mlx-4Bit: Quantization
- nabi-chan/Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-MLX-4bit: Quantization
- LibraxisAI/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-vmlx-mxfp4: Quantization
- blockblockblock/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-exl3-5.0bpw: Same weights, converted
- blockblockblock/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-exl3-3.0bpw: Same weights, converted
- porschefreak/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-mlx-4Bit: Quantization
- YuYu1015/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-int4-AutoRound: Quantization
- Stickmouse/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-Q3_K_M-GGUF: Quantization
- nabi-chan/Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-MLX-5bit: Quantization
- LibraxisAI/Huihui-Qwen3.6-35B-A3B-Claude-4.7-Opus-abliterated-vmlx-mxfp8: Quantization
7 other repos hold byte-for-byte identical weight files.