sasa2000/MiroThinker-1.7-mini-heretic
Abliterated vs miromind-ai/MiroThinker-1.7-mini. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 44 of 48 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/193, MLP 774/2613, Attention 37/192, lm_head 0/1
2 derivatives checked
- mradermacher/MiroThinker-1.7-mini-heretic-i1-GGUF: Quantization
- mradermacher/MiroThinker-1.7-mini-heretic-GGUF: Quantization