rbinrs/Qwen2.5-14B-Instruct-abliterated-v2
The rbinrs/Qwen2.5-14B-Instruct-abliterated-v2 is a 14.8 billion parameter instruction-tuned causal language model, based on the Qwen2.5 architecture, developed by huihui-ai. This model is an uncensored version of Qwen2.5-14B-Instruct, created using an 'abliteration' technique to remove safety alignments. With a context length of 32768 tokens, it is designed for applications requiring an instruction-following model without built-in content moderation.
Loading preview...
Model Overview
The rbinrs/Qwen2.5-14B-Instruct-abliterated-v2 is a 14.8 billion parameter instruction-tuned language model derived from the Qwen2.5-14B-Instruct base model. Developed by huihui-ai, this version has been specifically modified using an "abliteration" technique to remove inherent safety alignments and content moderation, resulting in an uncensored model.
Key Characteristics
- Base Model: Qwen/Qwen2.5-14B-Instruct.
- Parameter Count: 14.8 billion parameters.
- Context Length: Supports a context window of 32768 tokens.
- Uncensored Nature: Modified via abliteration to bypass typical safety filters, offering unrestricted content generation.
- Multilingual Support: Inherits multilingual capabilities from its base, supporting languages including Chinese, English, French, Spanish, German, and more.
Use Cases
This model is suitable for developers and researchers who require an instruction-following large language model without built-in content restrictions. It can be particularly useful for:
- Research into model safety and alignment bypasses.
- Applications requiring unfiltered text generation.
- Exploring the boundaries of LLM responses without predefined guardrails.
Usage and Availability
The model can be loaded and utilized with the Hugging Face transformers library. An Ollama version is also available for local deployment, simplifying access for various development environments. This v2 iteration represents an improvement over the previous Qwen2.5-14B-Instruct-abliterated version.