LRM-Conta-Detection-Arena/sft-conta-deepseek-distill-llama3-8b

Full fine-tune vs deepseek-ai/DeepSeek-R1-Distill-Llama-8B. Fine-tune: 93% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 0/1, Attention 128/128, MLP 96/96, Norms 14/65, lm_head 1/1

1 derivatives checked

• mradermacher/sft-conta-deepseek-distill-llama3-8b-GGUF: Same weights, converted

On the Hub · Downloads on Model Pulse