ArliAI/Qwen3.5-35B-A3B-Derestricted
Abliterated vs Qwen/Qwen3.5-35B-A3B. Looks like abliteration (refusal removal): only the output projections (down_proj, o_proj) changed, in 17 of 40 layers. Abliteration edits exactly these matrices to remove a "refusal direction"; everything else is untouched.
Tensors changed: MLP 34/1165, lm_head 0/1, Embeddings 0/5, Attention 4/392, other 0/1, Norms 0/247
5 derivatives checked
- mradermacher/Qwen3.5-35B-A3B-Derestricted-i1-GGUF: Quantization
- mradermacher/Qwen3.5-35B-A3B-Derestricted-GGUF: Quantization
- SufficientPrune3897/Qwen3.5-35B-A3B-Derestricted-Character-Creator-V2-GGUF: Full fine-tune
- jorge-erdb/Qwen3.5-35B-A3B-Derestricted-GGUF: Same weights, converted
- wolfie565656/Qwen3.5-35B-A3B-Derestricted-GGUF: Same weights, converted
1 other repos hold byte-for-byte identical weight files.