mergekit-community/mergekit-slerp-euzaldk

Full fine-tune vs openchat/openchat-3.5-0106. Fine-tune: 94% of the weights changed, across 30 of 32 layers.

Tensors changed: Norms 56/65, MLP 90/96, Attention 120/128

1 derivatives checked

On the Hub · Downloads on Model Pulse