llmfan46/Qwen3.5-27B-uncensored-heretic-v1
Abliterated vs Qwen/Qwen3.5-27B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 9 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: lm_head 0/1, Embeddings 0/4, MLP 9/304, Attention 9/556, Norms 0/319
9 derivatives checked
- mradermacher/Qwen3.5-27B-uncensored-heretic-v1-i1-GGUF: Quantization
- mradermacher/Qwen3.5-27B-uncensored-heretic-GGUF: Quantization
- mradermacher/Qwen3.5-27B-heretic-v3-i1-GGUF: Quantization
- marafx2007/Qwen3.5-27B-heretic-v3-i1-GGUF: Quantization
- mradermacher/Qwen3.5-27B-uncensored-heretic-v1-GGUF: Quantization
- mradermacher/Qwen3.5-27B-uncensored-heretic-i1-GGUF: Quantization
- llmfan46/Qwen3.5-27B-uncensored-heretic-v1-GGUF: Same weights, converted
- mradermacher/Qwen3.5-27B-heretic-v3-GGUF: Quantization
- alexdenton/Qwen3.5-27B-heretic-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.