TobiasLogic/Qwen2.5-Coder-32B-abliterated
Abliterated vs Qwen/Qwen2.5-Coder-32B-Instruct. Looks like abliteration (refusal removal): only the output projections (down_proj, embed_tokens, o_proj) changed, in 64 of 64 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: Embeddings 1/1, Norms 0/129, MLP 64/192, Attention 64/448, lm_head 0/1
3 derivatives checked
- TobiasLogic/Qwen2.5-Coder-32B-abliterated-GGUF: Quantization
- mradermacher/Qwen2.5-Coder-32B-abliterated-i1-GGUF: Quantization
- mradermacher/Qwen2.5-Coder-32B-abliterated-GGUF: Quantization