prithivMLmods/Qwen3.6-35B-A3B-Uncensored-Aggressive-GGUF
Image-Text-to-Text ā¢ 35B ā¢ Updated ā¢ 431 ā¢ 2
Paulo Faria
Azoresman
AI & ML interests
None yet
Recent Activity
liked a model about 3 hours ago
prithivMLmods/Qwen3.6-35B-A3B-Uncensored-Aggressive-GGUF liked a model about 3 hours ago
prithivMLmods/Qwen3.6-35B-A3B-Uncensored-Aggressive liked a model about 15 hours ago
OpenYourMind/OpenYourMind-Qwen3.6-35B-A3B-abliterated-uncensored-APEX-GGUFOrganizations
None yet
liked 2
models about 3 hours ago
liked a
model about 15 hours ago
reacted to SeaWolf-AI's post with š„ 9 days ago
Post
8687
š§¬ Introducing Darwin-9B-NEG ā the first model with Native Entropy Gating (NEG)
š Try it now: FINAL-Bench/Darwin-9B-NEG
š Q4 bit : FINAL-Bench/Darwin-9B-MFP4
We're thrilled to release Darwin-9B-NEG, a 9B-parameter reasoning model
that embeds an architecturally-internalised sense of self-confidence directly
into the transformer ā our proprietary Native Entropy Gating (NEG) technology.
š GPQA Diamond (198 PhD-level questions):
āø Baseline Darwin-9B (no NEG) ā 51.01 %
āø Pure NEG (greedy Ā· 1Ć cost) ā 63.64 % š„ +12.63 %p
āø + Permutation (4Ć cost) ā 76.26 %
āø + Ensemble Refinement (~20Ć) ā 84.34 % š
With only 9 billion parameters and 1Ć inference cost, Pure NEG jumps
+12.63 %p over the same model without NEG. Going all-in with ensemble
refinement pushes it to 84.34 % ā surpassing the published Qwen3.5-9B
leaderboard score (81.7 %) by +2.64 %p.
š¬ What makes NEG different from Multi-Turn Iteration (MTI)?
Classical MTI needs 3-8Ć extra inference passes. NEG instead lives
INSIDE the single decoding loop. Two tiny modules ride with the
transformer: NEG-Head predicts per-token entropy from the last hidden
state, and NEG-Gate conditionally restricts the top-k choice when
confidence is low. The gate activates in only 4.36 % of tokens ā
essentially free at inference time.
āØ Key differentiators
ā¢ Architecturally internalised ā model file *is* the feature
ā¢ 1Ć inference cost (vs. 3-8Ć for MTI)
ā¢ Drop-in with vLLM / SGLang / TGI / transformers ā no extra engine
ā¢ +12.63 %p reasoning at zero latency overhead
ā¢ Single-file deployment, Apache 2.0 licensed
š§¬ Lineage
Qwen/Qwen3.5-9B ā Darwin-9B-Opus (V7 evolutionary merge) ā Darwin-9B-NEG (V8 + NEG training)
#Darwin #NEG #NativeEntropyGating #GPQA #Reasoning #LLM #OpenSource #Apache2
š Try it now: FINAL-Bench/Darwin-9B-NEG
š Q4 bit : FINAL-Bench/Darwin-9B-MFP4
We're thrilled to release Darwin-9B-NEG, a 9B-parameter reasoning model
that embeds an architecturally-internalised sense of self-confidence directly
into the transformer ā our proprietary Native Entropy Gating (NEG) technology.
š GPQA Diamond (198 PhD-level questions):
āø Baseline Darwin-9B (no NEG) ā 51.01 %
āø Pure NEG (greedy Ā· 1Ć cost) ā 63.64 % š„ +12.63 %p
āø + Permutation (4Ć cost) ā 76.26 %
āø + Ensemble Refinement (~20Ć) ā 84.34 % š
With only 9 billion parameters and 1Ć inference cost, Pure NEG jumps
+12.63 %p over the same model without NEG. Going all-in with ensemble
refinement pushes it to 84.34 % ā surpassing the published Qwen3.5-9B
leaderboard score (81.7 %) by +2.64 %p.
š¬ What makes NEG different from Multi-Turn Iteration (MTI)?
Classical MTI needs 3-8Ć extra inference passes. NEG instead lives
INSIDE the single decoding loop. Two tiny modules ride with the
transformer: NEG-Head predicts per-token entropy from the last hidden
state, and NEG-Gate conditionally restricts the top-k choice when
confidence is low. The gate activates in only 4.36 % of tokens ā
essentially free at inference time.
āØ Key differentiators
ā¢ Architecturally internalised ā model file *is* the feature
ā¢ 1Ć inference cost (vs. 3-8Ć for MTI)
ā¢ Drop-in with vLLM / SGLang / TGI / transformers ā no extra engine
ā¢ +12.63 %p reasoning at zero latency overhead
ā¢ Single-file deployment, Apache 2.0 licensed
š§¬ Lineage
Qwen/Qwen3.5-9B ā Darwin-9B-Opus (V7 evolutionary merge) ā Darwin-9B-NEG (V8 + NEG training)
#Darwin #NEG #NativeEntropyGating #GPQA #Reasoning #LLM #OpenSource #Apache2
liked a
model 14 days ago
reacted to danielhanchen's post with š„ about 2 months ago
Post
3407
Introducing Unsloth Studio āØ
A new open-source web UI to train and run LLMs.
ā¢ Run models locally on Mac, Windows, Linux
ā¢ Train 500+ models 2x faster with 70% less VRAM
ā¢ Supports GGUF, vision, audio, embedding models
ā¢ Auto-create datasets from PDF, CSV, DOCX
ā¢ Self-healing tool calling and code execution
ā¢ Compare models side by side + export to GGUF
GitHub: https://github.com/unslothai/unsloth
Blog and Guide: https://unsloth.ai/docs/new/studio
Available now on Hugging Face, NVIDIA, Docker and Colab.
A new open-source web UI to train and run LLMs.
ā¢ Run models locally on Mac, Windows, Linux
ā¢ Train 500+ models 2x faster with 70% less VRAM
ā¢ Supports GGUF, vision, audio, embedding models
ā¢ Auto-create datasets from PDF, CSV, DOCX
ā¢ Self-healing tool calling and code execution
ā¢ Compare models side by side + export to GGUF
GitHub: https://github.com/unslothai/unsloth
Blog and Guide: https://unsloth.ai/docs/new/studio
Available now on Hugging Face, NVIDIA, Docker and Colab.
thinking mode help
5
#13 opened about 2 months ago
by 
wicesxs
liked a
model about 2 months ago
reacted to prithivMLmods's post with ā¤ļø about 2 months ago
Post
5049
The Qwen3.5 Multimodal Understanding Demo, powered by Qwen3.5-2B, is now available on HF Spaces! It is a lightweight model designed for fast image and video reasoning. Built with Gradio, the demo showcases Image QA, Video QA, object detection, and 2D point tracking, along with real-time token streaming.
š¤ Demo: prithivMLmods/Qwen-3.5-HF-Demo
ā Collection: https://huggingface.co/collections/prithivMLmods/multimodal-implementations
š Qwen3.5-2B: Qwen/Qwen3.5-2B
To learn more, visit the app page or the respective model pages.
š¤ Demo: prithivMLmods/Qwen-3.5-HF-Demo
ā Collection: https://huggingface.co/collections/prithivMLmods/multimodal-implementations
š Qwen3.5-2B: Qwen/Qwen3.5-2B
To learn more, visit the app page or the respective model pages.
liked 5
models about 2 months ago
llmfan46/Qwen3.5-35B-A3B-ultra-uncensored-heretic
Image-Text-to-Text ā¢ 35B ā¢ Updated ā¢ 1k ā¢ 31
HauhauCS/GLM-4.7-Flash-Uncensored-HauhauCS-Aggressive
30B ā¢ Updated ā¢ 8.77k ā¢ 26
HauhauCS/Qwen3.5-27B-Uncensored-HauhauCS-Aggressive
27B ā¢ Updated ā¢ 422k ā¢ 320
llmfan46/Qwen3.5-27B-ultra-uncensored-heretic-v1
Image-Text-to-Text ā¢ 27B ā¢ Updated ā¢ 806 ā¢ 27
llmfan46/Q3.5-BlueStar-27B-ultra-heretic
27B ā¢ Updated ā¢ 44 ā¢ 12
liked 2
models 2 months ago
Is `qwen3_nonthinking.jinja` available for disabling thinking?
7
#23 opened 2 months ago
by 
kraftDong
absolutely awesome
š 1
8
#1 opened 2 months ago
by 
baikw
liked a
model 2 months ago
Brain damage?
3
#2 opened 2 months ago
by 
St33lMouse