DeepSeek-V4.1-Flash-1017 Uncensored Model Trending on Hugging Face
Original title: AtomicChat/DeepSeek-V4.1-Flash-1017-Uncensored-GSQ-RCO-GGUF — new model trending #30 on Hugging Face
AtomicChat releases 763B DeepSeek-V4.1-Flash-1017 uncensored GGUF quantized for single 12GB GPU with Strata.
AtomicChat releases an uncensored version of DeepSeek-V4.1-Flash-1017 using GSQ-RCO quantization of routed experts. The model runs on a single 12GB GPU by streaming an Engram table from SSD. It eliminates refusals across topics while preserving task accuracy on benchmarks like GPQA-Diamond and AIME25.
- IQ2_XXS rung: 0.0% JailbreakBench vs 85.2% stock
- Reasoning benchmarks match stock (GPQA 83.3%, AIME 80.0% vs 76.7%)
- Edit only replaces dense residual writers; routed experts unchanged
- Requires 110GB Engram on fast SSD
- Vendors
- AtomicChat
- AI models
- DeepSeek-V4.1-Flash-1017
From the original article
3 min read at huggingface.cotext-generation · 17,384 downloads · 160 likes · license mit
By AtomicChat · huggingface.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.