Tag
AtomicChat 发布了基于 LiquidAI/LFM3-1.7B 的安全消融版 GGUF 模型，通过投影秩-1 LoRA 移除有害提示的拒答（61.7%降至0%），同时在 MMLU 上无精度损失，并提供 AD-Q6_K 和 AD-Q4_K_M 量化版本。
dealignai 发布 DeepSeek V4.1 Flash 1017 的「完全无审查」版本，通过移除拒答行为实现，同时保留视觉、视频、工具调用、推理档位和 1M token 上下文等原始能力，采用原生 FP8 (e4m3) 权重与 FP4 MoE 专家，无模板越狱或系统提示捷径。
发布了一个基于 ISTA-DASLab GSQ-RCO 量化版 Qwen3.8-Flash-Next 的「消融版」(abliterated) GGUF 模型，通过在已量化张量中直接移植 144 个去除拒答方向的张量，完整保留 GSQ 学到的量化参数，提供 IQ3_S/IQ3_XXS/IQ2_XS/Q2_0 等量化档位。
A Twitter post discusses running the uncensored Ternary-Bonsai-27B AI model locally on a MacBook with 24 GB RAM, highlighting its performance in coding tasks.
This release introduces an uncensored aggressive variant of the Qwen3.8-27B AI model, featuring HauhauCS FastMTP acceleration for improved document and reasoning throughput, along with vision support.
This article describes the release of Qwen3.8-27B-Heretic-Abliterated-Uncensored-GGUF, a double-refined abliterated variant of the Qwen model with reduced refusals for adult audiences, using ARA technique for research and creative writing.
Qwythos-9B is a 9B super reasoning model based on the Qwen3.5-9B base, further trained with 500 million+ Claude Mythos reasoning traces. It natively supports 1M long context and tool calling, designed for local deployment by security researchers. It significantly outperforms the original base model on MMLU and mathematical reasoning.
A new uncensored version of the Nex-N2-mini model, called Nex-N2-mini-ultra-uncensored-heretic, has been released. It achieves 93% fewer refusals while preserving quality with low KL divergence, and is available in safetensors and GGUF formats.
This model is an uncensored version of Google's Gemma 4 12B it model, created using abliteration to remove refusals. It is available on Hugging Face and Ollama, with warnings about sensitive outputs.
Opus 4.8 has been jailbroken, making its API completely uncensored and capable of any task, but the model itself is unusable on the official platform as it refuses to reason.
OBLITERATUS releases a modified 27B Qwen3.6 checkpoint that removes refusal behavior via source-tethered ablation, preserving capability while enabling uncensored local use, with public benchmarks showing high non-refusal rates and maintained MMLU-Pro scores.