UnfilteredAI/DAN-L3-R1-8B

Partial fine-tune vs deepseek-ai/DeepSeek-R1-Distill-Llama-8B. Partial change: 87% of the weights changed (attention, mlp, norms), in 32 of 32 layers.

Tensors changed: Embeddings 0/1, Attention 128/128, MLP 96/96, Norms 13/65, lm_head 0/1

2 derivatives checked

• mradermacher/DAN-L3-R1-8B-GGUF: Same weights, converted
• mradermacher/DAN-L3-R1-8B-i1-GGUF: Quantization

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse