Model Overview

nlpguy/Hermes-low-tune-2 is a 7 billion parameter language model developed by nlpguy. It was created using the task arithmetic merge method, combining multiple specialized models into a single, more versatile model. The base model for this merge is teknium/OpenHermes-2.5-Mistral-7B.

Merge Details

This model integrates capabilities from several OpenHermes-based models, including:

• simonveitner/Math-OpenHermes-2.5-Mistral-7B
• mlabonne/NeuralHermes-2.5-Mistral-7B-laser
• openaccess-ai-collective/dpopenhermes-alpha-v0
• mlabonne/NeuralHermes-2.5-Mistral-7B

Performance Highlights

Evaluated on the Open LLM Leaderboard, Hermes-low-tune-2 achieves an average score of 68.04. Notable scores include:

• AI2 Reasoning Challenge (25-Shot): 65.61
• HellaSwag (10-Shot): 84.47
• MMLU (5-Shot): 63.69
• GSM8k (5-Shot): 63.53

Use Cases

This model is suitable for general conversational AI applications, question answering, and tasks requiring a balance of reasoning and common sense, benefiting from the diverse strengths of its merged components.