HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive
LoRA-style vs Qwen/Qwen3.5-9B. HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive is a GGUF (BF16) of different weights. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, o_proj, up_proj changed, in 23 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, MLP 39/96, Attention 7/32, Norms 0/81
11 derivatives checked
- TheCluster/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-MLX-mxfp4: Quantization
- LEONW24/Qwen3.5-9B-Uncensored: Quantization (fewer layers)
- AIconjured/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-MTP-GGUF-NVFP4: Quantization
- TheCluster/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-MLX-mxfp8: Quantization
- nandukmelath/Qwen3.5-9B-Uncensored-nothink-GGUF: Quantization
- TheCluster/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-MLX-8bit: Quantization
- nuofang/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-GGUF: Quantization
- nDimensional/Qwen3.5-9B-Uncensored-Safetensors: Same weights, converted
- TheCluster/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-MLX-bf16: Same weights, converted
- suldanpashir/Qwen3.5-9B-Uncensored: Quantization (fewer layers)
- YoungHungGayGymBoy/Tiny-Darkllama3.2-1B-Instruct: Not comparable yet