MCult01/muse-gemma4e4b-heretic
Full fine-tune vs coder3101/gemma-4-E4B-it-heretic. Fine-tune: 100% of the weights changed, across 42 of 42 layers.
Tensors changed: MLP 606/606, Norms 502/502, other 261/261, Attention 755/755, Embeddings 6/6
2 derivatives checked
- mradermacher/muse-gemma4e4b-heretic-i1-GGUF: Quantization (fewer layers)
- mradermacher/muse-gemma4e4b-heretic-GGUF: Same weights, converted (fewer layers)
6 other repos hold byte-for-byte identical weight files.