huihui-ai/Huihui-Qwen3-VL-32B-Instruct-abliterated
Abliterated vs Qwen/Qwen3-VL-32B-Instruct. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 64 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/4, Norms 0/373, MLP 64/316, Attention 64/364, lm_head 0/1
4 derivatives checked
- mradermacher/Huihui-Qwen3-VL-32B-Instruct-abliterated-GGUF: Quantization (fewer layers)
- mradermacher/Huihui-Qwen3-VL-32B-Instruct-abliterated-i1-GGUF: Quantization (fewer layers)
- Heouzen/Huihui-Qwen3-VL-32B-Instruct-FP8-abliterated: Full fine-tune
- divinetribe/Huihui-Qwen3-VL-32B-Instruct-abliterated-4bit-mlx: Quantization
2 other repos hold byte-for-byte identical weight files.