ar0cket1/qwen3-30b-a3b-base-reasoning-sft-nemotron-math-v4-500m
Full fine-tune vs Qwen/Qwen3-30B-A3B-Base. Fine-tune: 94% of the weights changed, across 48 of 48 layers.
Tensors changed: Embeddings 0/1, Norms 186/193, MLP 2613/2613, Attention 192/192, lm_head 1/1