DebateLabKIT/Llama-3.1-Argunaut-1-8B-SPIN
Partial fine-tune vs DebateLabKIT/Llama-3.1-Argunaut-1-8B-SFT. Partial change: 31% of the weights changed (attention, lm_head, mlp, norms), in 28 of 32 layers.
Tensors changed: Embeddings 0/1, Norms 12/65, MLP 27/96, Attention 36/128, lm_head 1/1
3 derivatives checked
- mradermacher/Llama-3.1-Argunaut-1-8B-SPIN-i1-GGUF: Quantization
- mradermacher/Llama-3.1-Argunaut-1-8B-SPIN-GGUF: Same weights, converted
- DebateLabKIT/Llama-3.1-Argunaut-1-8B-HIRPO: Partial fine-tune