sahilchachra/LFM2.5-8B-A1B-Uncensored

Full fine-tune vs LiquidAI/LFM2.5-8B-A1B. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: MLP 16/44, Embeddings 1/2, other 36/36, Norms 0/60, Attention 24/24

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse