ncgc/statichh-pythia-1.4b-sft-bf16

Full fine-tune vs EleutherAI/pythia-1.4b. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: Embeddings 2/2, Norms 81/98, Attention 96/96, MLP 96/96

1 derivatives checked

• mradermacher/statichh-pythia-1.4b-sft-bf16-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse