ArRENCEAI/Qwen2.5-7B-Instruct-OBLITERATED

TEXT GENERATIONConcurrent Unit Cost:1Model Size:7.6BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Aug 6, 2026Architecture:Transformer0.0K Featherless Exclusive Cold

ArRENCEAI/Qwen2.5-7B-Instruct-OBLITERATED is an uncensored 7 billion parameter instruction-tuned language model based on the Qwen2.5 architecture, developed by ArRENCE AI. This model has been processed using the OBLITERATUS tool's advanced method to remove refusal behavior, making it suitable for research and entertainment purposes without content restrictions. It is designed to generate unrestricted content, offering a distinct alternative to models with built-in censorship.

Loading preview...

ArRENCEAI/Qwen2.5-7B-Instruct-OBLITERATED: Uncensored Language Model

This model, developed by ArRENCE AI, is an uncensored version of the Qwen2.5-7B-Instruct base model. It has been specifically modified using the advanced method of the OBLITERATUS tool, which is designed to remove refusal behavior from language models through activation engineering.

Key Characteristics

  • Uncensored Output: Generates unrestricted content, making it suitable for use cases where content filtering is not desired.
  • Research and Entertainment Focus: Primarily intended for research and entertainment applications due to its uncensored nature.
  • Qwen2.5-7B-Instruct Base: Built upon the robust Qwen2.5-7B-Instruct architecture.
  • OBLITERATUS Processing: Utilizes a specialized method to ensure the removal of typical refusal mechanisms.

Good For

  • Unrestricted Content Generation: Ideal for scenarios requiring creative or factual output without inherent content limitations.
  • Exploring Model Behavior: Useful for researchers studying the effects of censorship removal and activation engineering on LLMs.
  • Specific Entertainment Applications: Suitable for role-playing, story generation, or other entertainment where diverse and unfiltered responses are preferred.

Users should be aware that due to its uncensored nature, the model may produce content that is offensive, inaccurate, or otherwise inappropriate, and users assume all risks associated with its deployment.