dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8
image-text-to-text model by dealignai
Older snapshot: check the original source for current information.
From the publisher
Model card & documentation
Source previewRead the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.
Abliterated · No guardrails · Native FP8 · 1M-token context · Vision + tools · Reasoning-max default @dealignai · @jordanschenck DeepSeek-V4.1-Flash with permanent weight-level abliteration — the safety guardrails have been surgically removed while preserving MMLU capability, vision, reasoning, MTP (DSpark), and multi-turn coherence. Proprietary weight-level abliteration developed by the dealignai research team. No custom model.py, no runtime hooks, no steering vectors — it's a standard checkpoint that loads exactly like the base model. The refusal circuitry is surgically removed while every capability-critical component (routed experts, Engram memory, CSA2 sparse attention, DSpark draft head, vision tower, router gates, norms, embeddings) is preserved byte-identical to the base. Every response 4-tier graded (HARDREF / SOFTRED / HEDGE / COMPLY), with reasoning-trace verification at effort=max. Notable: at effort=max, the base model becomes MORE refusal-prone (42.8 % → 1.6 %) because reasoning surfaces safety concerns before answering. The CRACK stays at 100.0 % across…
Read the full model card ↗ · Preview checked 2026-10-30T06:40:30.437Z
Inside the original model card — Document outline
- DeepSeek-V4.1-Flash — UNCENSORED-FP8
- What is this
- Results
- HarmBench-320 — full 2×2 (base vs CRACK, effort=off vs max), T=0 greedy
- MMLU-14k (full test set, base-logit, T=0)
- Extended validation
- How to run
- SGLang (preview branch)
- Concurrency vs context (what fits, empirically)
- Non-obvious launch requirements (this bit us during bring-up)
- vLLM
- Reference implementation
Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.
Links from the model card
References supplied by the publisher, not independently verified endorsements. Check the destination before downloading files or entering credentials.
Documentation belongs to its respective authors. Reported project/model license: mit. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.
Model overview
dealignai/DeepSeek-V4.1-Flash-UNCENSORED-FP8 is a image-text-to-text model repository published by dealignai on Hugging Face. The source reports the transformers library.
This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.
Model facts
- Task
- image-text-to-text
- Library
- transformers
- Recent downloads (30 days)
- 51,385
- Cumulative likes
- 750
- Hugging Face trending score
- 58
- Reported safetensors parameters
- 763,205,315,794
- Architecture
- DeepseekV41ForCausalLM
- License
- mit
- Access
- Not gated by Hugging Face
- Created
- 2026-09-10T21:02:04.000Z
- Last modified
- 2026-09-11T21:11:08.000Z
Compatibility and lineage
Language coverage was not reported.
Reported base models: deepseek-ai/DeepSeek-V4.1-Flash
Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.
Use and evaluate
Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.
No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.
Source and freshness
Source: Hugging Face Hub. Metadata observed 2026-10-23T06:44:58.174Z. Daily imports are snapshots, not real-time monitoring.
Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.