llmfan46/Qwythos-9B-Claude-Mythos-5-1M-uncensored-heretic
Abliterated vs empero-ai/Qwythos-9B-Claude-Mythos-5-1M. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj, out_proj) changed, in 25 of 32 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: lm_head 0/1, Embeddings 0/4, Norms 0/215, Attention 25/332, MLP 22/208
4 derivatives checked
- llmfan46/Qwythos-9B-Claude-Mythos-5-1M-uncensored-heretic-GGUF: Same weights, converted
- fairy322/Qwythos-9B-Claude-Mythos-5-1M-uncensored-heretic-GGUF: Same weights, converted
- jauganaut1/Qwythos-9B-Claude-Mythos-5-1M-uncensored-heretic-GGUF: Same weights, converted
- Lunaunbanned/Qwythos-9B-Claude-Mythos-5-1M-uncensored-heretic-GGUF: Same weights, converted
1 other repos hold byte-for-byte identical weight files.