wangzhang/gpt-oss-120b-abliterated
Abliterated vs openai/gpt-oss-120b. Looks like abliteration (refusal removal): only the output projections (o_proj) changed, in 35 of 36 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: MLP 0/144, Norms 0/73, Attention 35/324, lm_head 0/1, Embeddings 0/1
2 derivatives checked
- mradermacher/gpt-oss-120b-abliterated-i1-GGUF: Quantization
- mradermacher/gpt-oss-120b-abliterated-GGUF: Quantization
2 other repos hold byte-for-byte identical weight files.