microsoft/HARC-Qwen2.5-7B-Instruct
HARC-Qwen2.5-7B-Instruct is a 7.6 billion parameter instruction-tuned causal language model developed by Microsoft, based on the Qwen2.5 architecture. This model integrates a HARC safety-alignment LoRA, making it specifically optimized for enhanced safety in conversational AI applications. It features a 32768 token context length, suitable for applications requiring robust and safe dialogue generation.
Loading preview...
HARC-Qwen2.5-7B-Instruct Overview
This model is a 7.6 billion parameter instruction-tuned language model, developed by Microsoft, that builds upon the Qwen2.5 architecture. Its primary distinguishing feature is the integration of a HARC (Human-Aligned Response Generation) safety-alignment LoRA, which has been merged into the base Qwen/Qwen2.5-7B-Instruct model. This makes it a standalone model specifically designed with enhanced safety protocols.
Key Capabilities
- Safety-Aligned Responses: Incorporates HARC safety-alignment for generating more secure and appropriate outputs.
- Instruction Following: Tuned to follow instructions effectively, making it suitable for various conversational tasks.
- Extended Context Window: Supports a context length of 32768 tokens, allowing for processing and generating longer sequences of text.
Good For
- Safe AI Applications: Ideal for use cases where generating human-aligned and safe responses is critical.
- Conversational Agents: Suitable for chatbots and virtual assistants that require robust instruction following and safety features.
- Research in AI Safety: Provides a foundation for exploring and developing safer large language models, as detailed in the associated HARC paper (arXiv:2607.00572).