Nikity/lille-130m-instruct

Full fine-tune vs Nikity/lille-130m-base. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: Norms 49/49, Attention 48/48, MLP 72/72, other 0/3, Embeddings 1/1

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

On the Hub · Downloads on Model Pulse