mradermacher/Qwen3-4B-BiasExpert-GGUF
Same weights, converted vs EmergentMethods/Qwen3-4B-BiasExpert. Same weights stored in another dtype (BF16 → F16/F32).
Tensors changed: Embeddings 0/1, Norms 0/145, MLP 0/108, Attention 0/144
Same weights, converted vs EmergentMethods/Qwen3-4B-BiasExpert. Same weights stored in another dtype (BF16 → F16/F32).
Tensors changed: Embeddings 0/1, Norms 0/145, MLP 0/108, Attention 0/144