neuqrui/EOPSA-Qwen3-1.7B
neuqrui/EOPSA-Qwen3-1.7B is a 2 billion parameter causal language model developed by neuqrui, based on the Qwen3-1.7B architecture. This model is specifically fine-tuned using the Efficient On-Policy Self-Distilled Safety Alignment (EOPSA) method. Its primary differentiator is its enhanced safety alignment, making it suitable for applications requiring robust content moderation and responsible AI interactions. The model supports a context length of 32768 tokens.
Loading preview...
EOPSA-Qwen3-1.7B Overview
This model, neuqrui/EOPSA-Qwen3-1.7B, is a 2 billion parameter language model derived from the Qwen3-1.7B architecture. It has been specifically fine-tuned by neuqrui using the Efficient On-Policy Self-Distilled Safety Alignment (EOPSA) method. This alignment process is designed to enhance the model's safety characteristics, making it more robust against generating harmful or undesirable content.
Key Capabilities
- Enhanced Safety Alignment: Utilizes the EOPSA method for improved content moderation and responsible AI behavior.
- Qwen3-1.7B Base: Benefits from the foundational capabilities of the Qwen3-1.7B model.
- Efficient Fine-tuning: Leverages an efficient on-policy self-distillation approach for safety training.
- Large Context Window: Supports a context length of 32768 tokens, allowing for processing longer inputs.
Good For
- Applications requiring a language model with a strong emphasis on safety and responsible output generation.
- Use cases where mitigating harmful content is a critical requirement.
- Developers looking for a Qwen3-based model with specialized safety fine-tuning.
Further details on the training methodology and safety evaluation can be found in the associated training code and safety evaluation repository.