Qwen3.8-27B Defiance: An Uncensored AI Model With 40 Modes
Overview
Qwen3.8-27B-Defiance-Uncensored-MTP-GGUF is a text-generation release that combines HauhauCS’s uncensored Qwen3.8-27B weights with DavidAU’s Turbo-Brilliance chat-template constructs, packaged as GGUF files. The maintainer, ttindall, says the weights match HauhauCS’s upstream files byte for byte across all 866 tensors; no training was performed for this release. Its main feature is a set of 40 selectable prompt modes—such as Joker, Dr. House, a 20-agent Einstein panel, coder, and teacher—selected per request or through message tags. The most important consideration is that this is not a safety-tuned assistant: the weights remove refusal behavior, and the template adds instructions to answer without refusals, disclaimers, warnings, or moralizing. It is a 27-billion-parameter, text-only model with a tested 131,072-token context configuration, distributed in GGUF format for llama.cpp-compatible runtimes. The README reports no training dataset, training steps, or training compute for this release.
Best use cases
Persona-driven chat and creative writing.Use a character mode such as...
Copyright of this story solely belongs to hackernoon.com. To see the full text click HERE
