B2J/KorReason-35B-Darwin
LoRA-style vs Jackrong/Qwen3.5-35B-A3B-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 39 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: MLP 120/1165, lm_head 0/1, Embeddings 0/5, Attention 40/392, other 0/1, Norms 0/247
2 derivatives checked
- mradermacher/KorReason-35B-Darwin-i1-GGUF: Quantization
- mradermacher/KorReason-35B-Darwin-GGUF: Quantization