NeelRajani/Qwen3-0.6B-Base_SFT-safety75_ADV-pku-v00.01
NeelRajani/Qwen3-0.6B-Base_SFT-safety75_ADV-pku-v00.01 is a 0.8 billion parameter Qwen3-based language model fine-tuned by NeelRajani. This model is specifically trained using Supervised Fine-Tuning (SFT) on the PKU-SafeRLHF_alpaca3-8b_severity-ge-2 dataset, focusing on safety and advanced instruction following. It is designed to enhance safety responses and instruction adherence, making it suitable for applications requiring robust content moderation and controlled output generation.
Loading preview...
Overview
This model, NeelRajani/Qwen3-0.6B-Base_SFT-safety75_ADV-pku-v00.01, is a 0.8 billion parameter variant of the Qwen3 architecture. It has been fine-tuned by NeelRajani from an existing Qwen3-0.6B-Base_SFT_safety model. The training specifically utilized the NeelRajani/PKU-SafeRLHF_alpaca3-8b_severity-ge-2 dataset, focusing on improving safety and instruction following capabilities.
Key Capabilities
- Enhanced Safety: The model is specifically trained to improve its responses in terms of safety, likely reducing the generation of harmful or inappropriate content.
- Advanced Instruction Following: Fine-tuning on the PKU-SafeRLHF dataset suggests an emphasis on adhering to complex instructions and generating more controlled outputs.
- Supervised Fine-Tuning (SFT): The model was trained using the SFT method with the TRL library, indicating a direct learning approach from high-quality, safety-aligned examples.
Good For
- Applications requiring a small, efficient language model with a strong emphasis on safety and controlled output.
- Use cases where adherence to specific instructions and avoidance of undesirable content generation are critical.
- Integration into systems that benefit from a model pre-aligned for safety, potentially reducing the need for extensive post-processing or filtering.