NexesMess/Llama_3.2_1b_Dolsun_Fusion

Full fine-tune vs Nexesenex/Dolphin3.0-Llama3.1-1B-abliterated. Fine-tune: 95% of the weights changed, across 16 of 16 layers.

Tensors changed: Embeddings 1/1, Norms 29/33, MLP 45/48, Attention 60/64

1 derivatives checked

• mradermacher/Llama_3.2_1b_Dolsun_Fusion-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse