lucaswychan/Nemotron-Research-Reasoning-Qwen-1.5B-Reasoning-Embedding

Full fine-tune vs nvidia/Nemotron-Research-Reasoning-Qwen-1.5B. Fine-tune: 100% of the weights changed, across 28 of 28 layers.

Tensors changed: Embeddings 1/1, Norms 57/57, MLP 84/84, Attention 196/196

1 derivatives checked