qikp/LFM2.5-230M-Thinking-v2

Full fine-tune vs LiquidAI/LFM2.5-230M-Base. Fine-tune: 100% of the weights changed, across 14 of 14 layers.

Tensors changed: Embeddings 2/2, other 24/24, MLP 42/42, Norms 36/40, Attention 24/24

1 derivatives checked