sert121/orpo_pythia2_hh_helpful

Full fine-tune vs EleutherAI/pythia-2.8b. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 2/2, Norms 130/130, Attention 128/128, MLP 128/128

1 derivatives checked

• mradermacher/orpo_pythia2_hh_helpful-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse