Hugging Face The huihui-ai/Qwen2.5-14B-Instruct-abliterated model is a 14.8 billion parameter instruction-tuned causal language model, derived from Qwen's Qwen2.5-14B-Instruct. This version has been specifically modified using abliteration techniques to remove censorship, making it suitable for applications requiring unfiltered responses. It maintains the original model's 131072 token context length, focusing on providing direct and uncensored conversational capabilities.
Overview
huihui-ai/Qwen2.5-14B-Instruct-abliterated is a 14.8 billion parameter instruction-tuned language model based on the Qwen2.5-14B-Instruct architecture. Its primary distinction lies in its "abliterated" nature, meaning it has undergone a process to remove inherent censorship present in the base model. This modification aims to provide users with an uncensored conversational AI experience, allowing for a broader range of responses without content filtering.
Key Capabilities
- Uncensored Responses: Designed to generate unfiltered content, bypassing typical safety and censorship mechanisms found in standard instruction-tuned models.
- Instruction Following: Retains the instruction-following capabilities of the original Qwen2.5-14B-Instruct model.
- Extended Context Window: Supports a substantial context length of 131072 tokens, enabling detailed and lengthy conversations.
Good For
- Use cases where unfiltered or uncensored text generation is explicitly required and acceptable.
- Research into model censorship and its removal techniques.
- Applications that need a powerful 14.8B parameter model with a very large context window, free from content restrictions.
Users should be aware that an updated version, Qwen2.5-14B-Instruct-abliterated-v2, is available and recommended for newer deployments.