saidutta69/Parable-Granite-4.1-3B-Claude-Fable-5-heretic
Abliterated vs AnkitAI/Parable-Granite-4.1-3B-Claude-Fable-5. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 25 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/81, MLP 3/120, Attention 25/160
2 derivatives checked
- mradermacher/Parable-Granite-4.1-3B-Claude-Fable-5-heretic-i1-GGUF: Quantization
- mradermacher/Parable-Granite-4.1-3B-Claude-Fable-5-heretic-GGUF: Same weights, converted