huihui-ai/Huihui-Qwen3-14B-abliterated-v2
Abliterated vs Qwen/Qwen3-14B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 40 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/161, MLP 40/120, Attention 40/160, lm_head 0/1
5 derivatives checked
- mradermacher/Huihui-Qwen3-14B-abliterated-v2-GGUF: Quantization
- mradermacher/Huihui-Qwen3-14B-abliterated-v2-i1-GGUF: Quantization
- warshanks/Huihui-Qwen3-14B-abliterated-v2-AWQ: Quantization
- ms122222/mein_qwen_14b_merged: LoRA-style
- pj1983/Huihui-Qwen3-14B-abliterated-v2-mlx-4bit: Quantization
1 other repos hold byte-for-byte identical weight files.