JackHsieh/Qwen3-4B-Instruct-2507.prestar-RL.reason-only.lr7e-7-kl0.step-2176

Full fine-tune vs Qwen/Qwen3-4B-Instruct-2507. Fine-tune: 90% of the weights changed, across 36 of 36 layers.

Tensors changed: Embeddings 0/1, Norms 101/145, MLP 108/108, Attention 144/144

On the Hub · Downloads on Model Pulse