Weyaxi/SauerkrautLM-UNA-SOLAR-Instruct
Weyaxi/SauerkrautLM-UNA-SOLAR-Instruct is a 10.7 billion parameter instruction-tuned causal language model created by Weyaxi, developed by merging VAGOsolutions/SauerkrautLM-SOLAR-Instruct and fblgit/UNA-SOLAR-10.7B-Instruct-v1.0 using mergekit. This model, with a 4096 token context length, achieved the top position on the Open LLM Leaderboard as of December 2023, demonstrating strong general reasoning and language understanding capabilities. It is optimized for diverse instruction-following tasks, making it suitable for a wide range of applications requiring robust conversational AI.
Loading preview...
SauerkrautLM-UNA-SOLAR-Instruct Overview
Weyaxi/SauerkrautLM-UNA-SOLAR-Instruct is a 10.7 billion parameter instruction-tuned language model. It was created by Weyaxi through a merge of VAGOsolutions/SauerkrautLM-SOLAR-Instruct and fblgit/UNA-SOLAR-10.7B-Instruct-v1.0 using the mergekit tool. This model notably secured the first-place position on the Open LLM Leaderboard as of December 2023, highlighting its strong performance across various benchmarks.
Key Capabilities & Performance
This model demonstrates robust performance in general language understanding and reasoning tasks, as indicated by its leaderboard scores:
- Average Score: 74.26
- AI2 Reasoning Challenge (25-Shot): 70.90
- HellaSwag (10-Shot): 88.30
- MMLU (5-Shot): 66.15
- TruthfulQA (0-shot): 71.80
- Winogrande (5-shot): 83.74
- GSM8k (5-shot): 64.67
Use Cases
Given its strong benchmark performance and instruction-tuned nature, SauerkrautLM-UNA-SOLAR-Instruct is well-suited for:
- General-purpose conversational AI.
- Tasks requiring robust reasoning and problem-solving.
- Applications benefiting from a model with high accuracy on common language understanding benchmarks.
Quantized versions (GPTQ, GGUF, AWQ) are available from TheBloke for optimized deployment.