mlabonne/Qwen3-0.6B-abliterated
Abliterated vs Qwen/Qwen3-0.6B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 28 of 28 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/113, MLP 28/84, Attention 28/112
3 derivatives checked
- bartowski/mlabonne_Qwen3-0.6B-abliterated-GGUF: Abliterated
- mradermacher/Qwen3-0.6B-abliterated-GGUF: Abliterated
- jaahas/Qwen3-0.6B-abliterated-Q4_K_M-GGUF: Quantization
2 other repos hold byte-for-byte identical weight files.