vishnukv/WestSeverusJaskier
vishnukv/WestSeverusJaskier is a 7 billion parameter language model created by merging PetroGPT/WestSeverus-7B-DPO and bardsai/jaskier-7b-dpo-v6.1 using the SLERP method. This merged model demonstrates strong general reasoning capabilities, achieving an average score of 75.67 on the Open LLM Leaderboard across various benchmarks. It is suitable for tasks requiring robust language understanding and generation, particularly where a blend of its constituent models' strengths is beneficial.
Loading preview...
Model Overview
vishnukv/WestSeverusJaskier is a 7 billion parameter language model resulting from a merge of two pre-trained models: PetroGPT/WestSeverus-7B-DPO and bardsai/jaskier-7b-dpo-v6.1. This merge was performed using the SLERP (Spherical Linear Interpolation) method, aiming to combine the strengths of its base models.
Performance Highlights
Evaluated on the Open LLM Leaderboard, WestSeverusJaskier achieved an average score of 75.67. Specific benchmark results include:
- AI2 Reasoning Challenge (25-Shot): 71.76
- HellaSwag (10-Shot): 88.16
- MMLU (5-Shot): 64.94
- TruthfulQA (0-shot): 73.18
- Winogrande (5-shot): 82.87
- GSM8k (5-shot): 73.09
These scores indicate a well-rounded performance across various reasoning, common sense, and language understanding tasks.
Ideal Use Cases
- General-purpose language generation: Leveraging the combined capabilities of its base models.
- Reasoning tasks: Demonstrated proficiency in benchmarks like AI2 Reasoning Challenge and GSM8k.
- Applications requiring robust language understanding: Suitable for tasks where strong performance across multiple metrics is desired.