DavidAU/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored
MoE built from dense models vs Jackrong/gpt-oss-120b-Distill-Qwen3-4B-Thinking. DavidAU/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored is a mixture of 12 experts with the same shape as Jackrong/gpt-oss-120b-Distill-Qwen3-4B-Thinking's MLP, but none of them is a copy of it: they come from other models or were trained further. Shared tensors (attention, norms, embeddings): 35 of 290 identical.
6 derivatives checked
- DavidAU/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored-GGUF: Quantization
- mradermacher/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored-i1-GGUF: Quantization
- mradermacher/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored-GGUF: Quantization
- mxderncultxre/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored-GGUF: Quantization
- MotoMadMobile/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored-GGUF: Quantization
- advanubis/Qwen3-48B-A4B-Savant-Commander-Distill-12X-Closed-Open-Heretic-Uncensored-GGUF: Quantization
2 other repos hold byte-for-byte identical weight files.