HauhauCS/Qwen3.5-35B-A3B-Uncensored-HauhauCS-Aggressive
LoRA-style vs Qwen/Qwen3.5-35B-A3B. HauhauCS/Qwen3.5-35B-A3B-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, shared_expert_gate, up_proj changed, in 31 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Attention 7/40, Norms 0/101, MLP 98/200
5 derivatives checked
- TheCluster/Qwen3.5-35B-A3B-Uncensored-HauhauCS-Aggressive-MLX-mxfp4: Quantization
- cahlen/Qwen3.5-35B-A3B-Uncensored-HauhauCS-Aggressive-GGUF: Quantization
- froggeric/Qwen3.5-35B-A3B-Uncensored-FernflowerAI-MLX-8bit: Quantization
- TheCluster/Qwen3.5-35B-A3B-Uncensored-HauhauCS-Aggressive-MLX-mxfp8: Quantization
- froggeric/Qwen3.5-35B-A3B-Uncensored-FernflowerAI-MLX-4bit: Quantization