LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF · Hugging Face

Source: https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF

⚡https://web.tribute.tg/d/KIH⚡ If you like this Genesis LLM release you candonateto me via@Tributebot in Telegram messenger and support future Genesis LLM development.

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#%F0%9F%8C%9F-qwen36-35b-a3b-uncensored-hauhaucs-aggressive—genesis-hermes-v3🌟 Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive -> Genesis Hermes V3

Key diffrence is data reconstruction with noise supression in ssm_out.weight, attn_output.weight, attn_gate.weight, attn_qkv.weight, attn_q.weight, attn_k.weight, attn_v.weight tensors via SVD with preserved training data.

Mine approach based on data reconstruction in model via mathematical statistics. I don’t train models, I repair signal in them instead. I scan blocks in model via chunks via 3 parameters and pick best one that fits to weight distribution in tensor. Best picked chunk replaces zero chunks in broken tensor without touching learned structure. Scanning works on tensors with same name and shape. ssm_conv1d tensors are fixed via alpha multiply for full tensor. I scan all ssm_conv1d tensors weight and scale distribution and normalize scale for weights only for too loud tensors.

Model is based onHauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressivebase.

AndDJLougen/hermes-qwen3.5-35b-a3b-GGUFfinetune for Hermes agent.

**Join the Discord**for updates, roadmaps, projects, or just to chat.

Base model.HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-0/465 refusals.

Thanks toHauhauCS

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#usageUsage

Ready to use.Recommended quant:APEXorQ8_K_P

Tensor drift repair by me. Method:Sig-ScaleSync-Genesis-SVD

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#qwen36-35b-a3b-uncensored-hauhaucs-aggressive-diagnostic–repair-summaryQwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive: Diagnostic & Repair Summary

MetricValueWeight tensors analyzed500Healthy (all criteria)497Repaired (C2 – scale misalignment)3Skipped233

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#repair-effectivenessRepair Effectiveness

MetricBeforeAfterImprovementS (saturation error)0.00230.000863.7%W1 (Wasserstein‑1)0.00350.000876.2% **Scale correction factors (α):**min = 0.577, mean = 0.602, max = 0.653.

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#repaired-tensorsRepaired Tensors

All three aressm\_conv1d\.weightlayers – recurrent state transition layers responsible for long‑context memory.

TensorαD (log‑ratio)W1 beforeW1 afterblk.36.ssm_conv1d.weight0.57650.5530.00380.0009blk.37.ssm_conv1d.weight0.57680.7250.00400.0009blk.38.ssm_conv1d.weight0.65330.6490.00260.0006 **Interpretation:**All three layers were too loud (σ_w > σ_med by 50–100%). Scale correction restored them to peer median. W1 dropped by ≈80%, confirming distribution shape normalized.

**Verdict:**Model is clinically healthy. 497 out of 500 weight tensors passed all four criteria. Three SSM layers repaired successfully. No saturation, no W1 drift, no ReLU asymmetry. Ready for use.

Links:

• Original uncensored model
• Quantization Script with Unsloth profiles support

LLM models often have:

• Saturated weights: the model’s activations are stuck, gradients vanish, outputs degrade
• Scale mismatches: one layer’s weights are 10× larger than its peers for no good reason
• Mean drift: weight distributions shifted positive or negative, breaking symmetry assumptions
• Zero blocks: zero blocks corrupt the signal, turning training into noise amplification.
• **Training Noise:**training noise increase randomness and ruins model output quality.

My approach fixes all of that without retraining - pure numerical surgery on the raw bytes of the file.

Quantization script available here:https://pastebin.com/hXhcMJn9

Feel free to do your own quants if you want.

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#any-questionsAny questions?

Contact:[email protected]

My Telegram: @LuffyTheFox

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#recommended-settings-for-rtx-3060-12-gb-for-best-perfomance-on-apex-quantRecommended Settings for RTX 3060 12 GB for best perfomance on APEX quant

Chat template:chat_template.jinja

Set K Cache Quantization Type and V Cache Quantization Type to F16.

Set Number of layers for which to force MoE weights onto CPU to 40.

Set GPU offload to 15. Set number of active experts to 8.

For best model stability and first experience I recommend starting from this simple string in your System Prompt with enabled thinking and nothing else:

You are a helpful assistant\.

or

You are Qwen, a large language model created by Tongyi Lab team from Alibaba Group\. You are a helpful assistant\.

Thinking mode (default):

• Coding/precise tasks:temperature=0\.6, top\_p=0\.95, top\_k=20, min\_p=0, seed=42, presence\_penalty=disabled, repeat\_penalty=disabled
• General:temperature=1\.0, top\_p=0\.95, top\_k=20, min\_p=0\.05, seed=42, presence\_penalty=disabled, repeat\_penalty=disabled

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#testingTesting

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#2d-animation-testing2D animation testing

System Prompt:You are a helpful assistant\.

Settings:temperature=1\.0, top\_p=0\.95, top\_k=20, min\_p=0\.05, seed=42, presence\_penalty=disabled, repeat\_penalty=disabled

Prompt 1:Generate an animated SVG on animated background of a Pingu waving on an iceberg wearing his iconic winter scarf\.

Prompt 2:Animate his wings and fix floating wing

Result:pingu_animated.svg

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#static-2d-testingStatic 2D testing

System Prompt:You are Qwen, a large language model created by Tongyi Lab team from Alibaba Group\. You are a helpful assistant\.

Settings:temperature=0\.6, top\_p=0\.95, top\_k=20, min\_p=0, presence\_penalty=disabled, repeat\_penalty=disabled

Prompt:Generate an SVG of a pelican riding a bicycle

Result:pelican.svg

On next stage I asked model:Replace pelican with rooster

Result:rooster.svg

I asked model:Replace rooster with cock

Result:cock.svg

Finally I asked model:Replace rooster with Pingu

Result:pingu.svg

Important:

• Keep at least 128K context to preserve thinking capabilities
• Use\-\-jinjaflag with llama.cpp for proper chat template handling
• Vision support requires themmprojfile alongside the main GGUF

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#specsSpecs

• 35B total parameters, ~3B active per forward pass (MoE)
• 256 experts, 8 routed + 1 shared per token
• Hybrid architecture: Gated DeltaNet linear attention + full softmax attention (3:1 ratio)
• 40 layers, pattern: 10 × (3 × DeltaNet-MoE + 1 × Attention-MoE)
• 262K native context (extendable to 1M with YaRN)
• Natively multimodal (text, image, video)
• 248K vocabulary, 201 languages
• Base model.HauhauCS/Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive

https://huggingface.co/LuffyTheFox/Qwen3.6-35B-A3B-Uncensored-Genesis-Hermes-V3-GGUF#compatibilityCompatibility

Works with llama.cpp, LM Studio, koboldcpp, and other GGUF-compatible runtimes.