richardyoung/Olmo-3-7B-Think-heretic
Abliterated vs allenai/Olmo-3-7B-Think. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 20 of 32 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, MLP 17/96, Norms 0/129, Attention 18/128, lm_head 0/1
3 derivatives checked
- mradermacher/Olmo-3-7B-Think-heretic-i1-GGUF: Quantization
- mradermacher/Olmo-3-7B-Think-heretic-GGUF: Same weights, converted
- richardyoung/Olmo-3-7B-Think-heretic-GGUF: Quantization