Weyaxi/GenAI-Nova-13B
Weyaxi/GenAI-Nova-13B is a 13 billion parameter language model developed by PulsarAI, evaluated on the Open LLM Leaderboard. It demonstrates strong performance across various benchmarks, including 83.27 on HellaSwag and 77.35 on Winogrande. This model is suitable for general language understanding and generation tasks, particularly those requiring robust common sense reasoning and factual recall.
Loading preview...
Overview
Weyaxi/GenAI-Nova-13B is a 13 billion parameter language model, developed by PulsarAI, that has been evaluated on the Hugging Face Open LLM Leaderboard. This model demonstrates a balanced performance across a range of benchmarks, indicating its capability for diverse natural language processing tasks.
Key Capabilities
- General Language Understanding: Achieves an average score of 51.53 across the evaluated benchmarks.
- Common Sense Reasoning: Scores 83.27 on HellaSwag (10-shot) and 77.35 on Winogrande (5-shot), highlighting its ability to handle everyday reasoning.
- Factual Recall: Demonstrates a MMLU (5-shot) score of 59.47, indicating proficiency in multi-task language understanding.
- Question Answering: Achieves 62.29 on ARC (25-shot) and 51.79 on TruthfulQA (0-shot), showcasing its capacity for accurate information retrieval and generation.
Performance Highlights
| Metric | Value |
|---|---|
| Avg. | 51.53 |
| ARC (25-shot) | 62.29 |
| HellaSwag (10-shot) | 83.27 |
| MMLU (5-shot) | 59.47 |
| TruthfulQA (0-shot) | 51.79 |
| Winogrande (5-shot) | 77.35 |
| GSM8K (5-shot) | 7.73 |
| DROP (3-shot) | 18.82 |
Good For
This model is well-suited for applications requiring general-purpose language understanding and generation, where a balance of common sense, factual knowledge, and reasoning abilities is important. Its performance on benchmarks like HellaSwag and Winogrande suggests it can be effectively used for tasks involving contextual understanding and logical inference.