rodrigomt/Qwen3.5-4B-Uncensored-Aggressive

Partial fine-tune vs Qwen/Qwen3.5-4B. Partial change: 19% of the weights changed (attention, mlp, norms), in 28 of 32 layers.

Tensors changed: Embeddings 0/5, MLP 32/199, Attention 34/324, Norms 27/209, other 0/1

2 derivatives checked

• tinyopsec/Qwen3.5-4B-Uncensored-Aggressive-GGUF: Same weights, converted
• heavry/WordStitch-4B: Adapter

On the Hub · Downloads on Model Pulse