玩模型开放模型 · 清晰来源 · 自主选择
返回模型目录
HauhauCS /

Qwen3.5-2B-Uncensored-HauhauCS-Aggressive

HF 源站

HauhauCS/Qwen3.5-2B-Uncensored-HauhauCS-Aggressive is an uncensored open-source model hosted on Hugging Face.

GGUFChineseEnglishMultilingualLicense: apache-2.02Buncensoredqwen3.5qwenconversational
README.md说明 / 使用方法

license: apache-2.0 tags:

  • uncensored
  • qwen3.5
  • qwen
  • gguf language:
  • en
  • zh
  • multilingual

Qwen3.5-2B-Uncensored-HauhauCS-Aggressive

加入 Discord 获取最新更新、开发路线图、项目进展或参与交流讨论。

由 HauhauCS 去除安全限制(Uncensored)的 Qwen3.5-2B 模型。

关于本模型 (About)

0/465 次拒绝(Refusal)。 完全解除安全限制,且能力零损耗。

未对数据集或模型能力进行删减更改。功能完备,100% 保留原始作者所预期的表现 —— 仅仅移除了拒绝回答的限制机制。

旨在打造市面上最优秀的无损去审查(lossless uncensored)模型。

激进版本说明 (Aggressive Variant)

具备更强力度的去审查处理,更为彻底地清除了拒绝机制。如果该版本在您的使用场景中显得过于宽松不受控,后续可能会推出 Balanced(平衡)版本。

注意: 该模型已完全解锁,不会拒绝用户的提示词指令。不过,它偶尔可能会在回答结尾附带一句简短的免责声明(例如“This is general information, not legal advice...”)。这是基座模型训练中自带的特性,并非模型拒绝回答 —— 实际的内容依然会完整生成。

下载列表 (Downloads)

文件 (File)量化类型 (Quant)大小 (Size)
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-BF16.ggufBF163.6 GB
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-Q8_0.ggufQ8_01.9 GB
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-Q6_K.ggufQ6_K1.5 GB
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-Q4_K_M.ggufQ4_K_M1.2 GB
mmproj-Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-f16.gguf视觉编码器 (Vision encoder)638 MB

视觉支持 (Vision support): 本模型原生支持多模态。mmproj 文件为视觉编码器 —— 如需使用图像/视频输入,您需要将其与主 GGUF 文件一同配合使用。请在 llama.cpp、LM Studio 或任何兼容的运行时中同时加载这两个文件。

模型规格 (Specs)

  • 2B 稠密参数量,24 层
  • 混合架构:门控 DeltaNet 线性注意力 (Gated DeltaNet linear attention) + 完整 softmax 注意力 (比例为 3:1)
  • 262K 原生上下文窗口(通过 YaRN 可扩展至 1M)
  • 原生支持多模态(文本、图像、视频)
  • 支持多 Token 预测 (MTP)
  • 248K 词表大小,覆盖 201 种语言
  • 基于 Qwen3.5-2B 构建

推荐配置参数 (Recommended Settings)

来自 Qwen 官方团队的推荐设置:

思考模式(默认,Thinking mode):

  • temperature=0.6, top_p=0.95, top_k=20, min_p=0

非思考模式(Non-thinking mode):

  • temperature=0.7, top_p=0.8, top_k=20, min_p=0

重要提示:

  • 请保持至少 128K 的上下文窗口,以保证模型的深度思考能力
  • 用于生产环境或高吞吐量场景时:建议使用 vLLM、SGLang 或 KTransformers

注意: 这是一个全新的架构(发布于 2026-03-02)。llama.cpp 的支持才刚上线不久 —— 请确保您使用的是最新的构建版本。兼容 llama.cpp、LM Studio、Jan、koboldcpp 等工具。

另可参考 27B 版本、9B 版本,以及 HauhauCS 发布的所有模型。

使用方法 (Usage)

适用于 llama.cpp、LM Studio、Jan、koboldcpp 等框架及推理工具。

源站原文

license: apache-2.0 tags:

  • uncensored
  • qwen3.5
  • qwen
  • gguf language:
  • en
  • zh
  • multilingual

Qwen3.5-2B-Uncensored-HauhauCS-Aggressive

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

Qwen3.5-2B uncensored by HauhauCS.

About

0/465 refusals. Fully uncensored with zero capability loss.

No changes to datasets or capabilities. Fully functional, 100% of what the original authors intended - just without the refusals.

These are meant to be the best lossless uncensored models out there.

Aggressive Variant

Stronger uncensoring with more thorough refusal removal. If this variant is too loose for your use case, a Balanced variant may follow.

Note: The model is fully unlocked and will not refuse prompts. However, it may occasionally append a short disclaimer at the end of a response (e.g. "This is general information, not legal advice..."). This is baked into the base model's training and not a refusal — the actual content is still generated in full.

Downloads

FileQuantSize
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-BF16.ggufBF163.6 GB
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-Q8_0.ggufQ8_01.9 GB
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-Q6_K.ggufQ6_K1.5 GB
Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-Q4_K_M.ggufQ4_K_M1.2 GB
mmproj-Qwen3.5-2B-Uncensored-HauhauCS-Aggressive-f16.ggufVision encoder638 MB

Vision support: This model is natively multimodal. The mmproj file is the vision encoder — you need it alongside the main GGUF to use image/video inputs. Load both files in llama.cpp, LM Studio, or any compatible runtime.

Specs

  • 2B dense parameters, 24 layers
  • Hybrid architecture: Gated DeltaNet linear attention + full softmax attention (3:1 ratio)
  • 262K native context (extendable to 1M with YaRN)
  • Natively multimodal (text, image, video)
  • Multi-token prediction (MTP) support
  • 248K vocabulary, 201 languages
  • Based on Qwen3.5-2B

Recommended Settings

From the official Qwen authors:

Thinking mode (default):

  • temperature=0.6, top_p=0.95, top_k=20, min_p=0

Non-thinking mode:

  • temperature=0.7, top_p=0.8, top_k=20, min_p=0

Important:

  • Maintain at least 128K context to preserve thinking capabilities
  • For production/high-throughput: use vLLM, SGLang, or KTransformers

Note: This is a brand new architecture (released 2026-03-02). llama.cpp support landed very recently — make sure you're on a recent build. Works with llama.cpp, LM Studio, Jan, koboldcpp, etc.

Also check out the 27B variant, the 9B variant, and all releases at HauhauCS.

Usage

Works with llama.cpp, LM Studio, Jan, koboldcpp, etc.

HUGGING FACE

上游数据

HF 下载量6.9万
HF 点赞量135

HF 下载量为源站最近 30 天统计,与本站点击分开。

SITE STATS

本站统计

下载点击—
请求数—
匿名去重—最近 7 天—

本站下载量记录按钮点击,不代表下载完成;短时连点会去重。

来源与版本

作者
HauhauCS
最近更新
2026/04/05
最新创建
2026/03/05
HF 趋势
237
数据快照
2026/09/16
版本
未知
许可证
apache-2.0