MultiverseComputingCAI/Qwen3-Next-80B-A3B-Thinking-Uncensored
Partial fine-tune vs Qwen/Qwen3-Next-80B-A3B-Thinking. Partial change: 6% of the weights changed (mlp, norms), in 40 of 48 layers.
Tensors changed: Embeddings 0/1, Norms 36/157, Attention 0/264, MLP 270/2514, lm_head 0/1
2 derivatives checked
- McG-221/Qwen3-Next-80B-A3B-Thinking-Uncensored-mlx-4Bit: Quantization
- Wwayu/Qwen3-Next-80B-A3B-Thinking-Uncensored-mlx-6Bit: Quantization
1 other repos hold byte-for-byte identical weight files.