SpaceTimee/Suri-Qwen-3.5-4B-Uncensored

Partial fine-tune vs Qwen/Qwen3.5-4B. Partial change: 56% of the weights changed (attention, mlp, norms), in 32 of 32 layers.

Tensors changed: Embeddings 0/4, MLP 96/196, Attention 32/320, Norms 24/203

1 derivatives checked

• SpaceTimee/Suri-Qwen-3.5-4B-Uncensored-GGUF: Quantization

On the Hub · Downloads on Model Pulse