seyf1elislam/WestKunai-Hermes-7b
seyf1elislam/WestKunai-Hermes-7b is a 7 billion parameter language model created by seyf1elislam, built upon the Mistral-7B-v0.1 architecture. This model is a merge of saishf/West-Hermes-7B and seyf1elislam/KunaiBeagle-Hermes-7b using the DARE TIES method, designed to combine the strengths of its constituent models. It achieves an average score of 73.51 on the Open LLM Leaderboard, indicating strong general language understanding and reasoning capabilities.
Loading preview...
Overview
seyf1elislam/WestKunai-Hermes-7b is a 7 billion parameter language model developed by seyf1elislam. It is a merged model, combining the capabilities of saishf/West-Hermes-7B and seyf1elislam/KunaiBeagle-Hermes-7b using the DARE TIES merge method. The base model for this merge was mistralai/Mistral-7B-v0.1, leveraging its robust foundation.
Key Capabilities
- General Language Understanding: Achieves an average score of 73.51 on the Open LLM Leaderboard, demonstrating proficiency across various tasks.
- Reasoning: Scores 71.16 on the AI2 Reasoning Challenge (25-Shot).
- Common Sense & World Knowledge: Performs well on HellaSwag (10-Shot) with 87.76 and Winogrande (5-shot) with 83.03.
- Mathematical Reasoning: Achieves 69.07 on GSM8k (5-shot).
Good For
- Applications requiring a balanced performance across general language tasks and reasoning.
- Developers looking for a 7B parameter model with a strong foundation from Mistral and enhanced capabilities from its merged components.
- Use cases where a model with competitive benchmark scores on the Open LLM Leaderboard is desired.