FINAL-Bench/Darwin-4B-Opus
LoRA-style vs arsovskidev/Gemma-4-E4B-Claude-4.6-Opus-Reasoning-Distilled. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 42 of 42 layers; norms, biases and embeddings are untouched.
Tensors changed: MLP 126/606, Norms 0/502, other 0/261, Attention 168/755, Embeddings 0/6
2 derivatives checked
- mradermacher/Darwin-4B-Opus-i1-GGUF: Quantization (fewer layers)
- mradermacher/Darwin-4B-Opus-GGUF: Same weights, converted (fewer layers)
3 other repos hold byte-for-byte identical weight files.