ChiKoi7/GPT-5-Distill-Qwen3-4B-Instruct-Heretic
Abliterated vs Jackrong/GPT-5-Distill-Qwen3-4B-Instruct. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 29 of 36 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 0/1, Norms 0/145, MLP 25/108, Attention 29/144
2 derivatives checked
- mradermacher/GPT-5-Distill-Qwen3-4B-Instruct-Heretic-i1-GGUF: Quantization
- mradermacher/GPT-5-Distill-Qwen3-4B-Instruct-Heretic-GGUF: Same weights, converted
1 other repos hold byte-for-byte identical weight files.