huihui-ai/Huihui-Qwen3.8-Flash-Next-abliterated
Abliterated vs Qwen/Qwen3.8-Flash-Next. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 48 of 48 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Norms 0/302, other 0/8, Attention 48/608, MLP 96/602, Embeddings 0/137, lm_head 0/1
7 derivatives checked
- mradermacher/Huihui-Qwen3.8-Flash-Next-abliterated-i1-GGUF: Not comparable yet
- mradermacher/Huihui-Qwen3.8-Flash-Next-abliterated-GGUF: Not comparable yet
- wy2pay/Qwen3.8-Flash-Next-abliterated-GGUF-IQ4XS: Not comparable yet
- Maximilian228/Huihui-Qwen3.8-Flash-Next-abliterated-NVFP4-GPTQ-Strata: Same weights, repacked (fewer layers)
- edp1096/Huihui-Qwen3.8-Flash-Next-abliterated-NVFP4-QAD: Quantization
- alesha-pro/Huihui-Qwen3.8-Flash-Next-abliterated-exl3-3bit-hq_h6_ng6: Same weights, converted
- alesha-pro/Huihui-Qwen3.8-Flash-Next-abliterated-exl3-4bit-hq_h6_ng6: Same weights, converted
1 other repos hold byte-for-byte identical weight files.