llmfan46/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved
Abliterated vs Qwen/Qwen3.6-35B-A3B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 34 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/5, Attention 34/392, MLP 60/399, Norms 0/247, lm_head 0/1, other 0/1
18 derivatives checked
- SC117/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-APEX-GGUF: Quantization
- llmfan46/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-GGUF: Same weights, converted
- llmfan46/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-GPTQ-Int4: Quantization
- mradermacher/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-i1-GGUF: Quantization
- SassyDiffusion/Qwen3.6-35B-A3B-heretic-MTP-GGUF: Same weights, converted
- mradermacher/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-GGUF: Quantization
- llmfan46/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-NVFP4-Experts-Only-GGUF: Quantization
- llmfan46/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-NVFP4-Experts-Only: Quantization
- Ikerbilbao/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-GGUF: Same weights, converted
- cvgro/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-APEX-GGUF: Quantization
- TopherAU/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-GGUF: Quantization
- hawhyhb/Qwen3.6-35B-A3B-Uncensored-Heretic-MTPLX-4bit-FP16: Quantization
- sleepy186247/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-mlx-8Bit: Quantization
- underlotus/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-oQ4e-mtp: Quantization
- sleepy186247/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-mlx-fp16: Partial fine-tune
- otherwhere1/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-mlx-8Bit: Quantization
- dopaemon/Qwen3.6-35B-A3B-uncensored-heretic-Native-MTP-Preserved-oQ4e-mtp: Quantization
- trigger2k25/Qwen3.6-35B-A3B-heretic-BF16-mtp: Partial fine-tune