siddharthsingh18/Qwen3-1.7B-KnK6RL-Polaris-GRPO-step900-bestpass1
Full fine-tune vs Qwen/Qwen3-1.7B-Base. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 97/113, MLP 84/84, Attention 112/112
Full fine-tune vs Qwen/Qwen3-1.7B-Base. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 97/113, MLP 84/84, Attention 112/112