NekoPunchBBB/Llama2-13b-hf-Open-Platypus-QLoRA-att
NekoPunchBBB/Llama2-13b-hf-Open-Platypus-QLoRA-att is a Llama 2-based language model, fine-tuned using QLoRA on the Open-Platypus dataset. This model demonstrates an average score of 47.33 on the Open LLM Leaderboard, with notable performance in HellaSwag (82.14) and ARC (58.87). It is suitable for general language understanding and generation tasks, particularly those benefiting from its fine-tuning on diverse datasets.
Loading preview...
Model Overview
NekoPunchBBB/Llama2-13b-hf-Open-Platypus-QLoRA-att is a Llama 2-based language model that has been fine-tuned using the QLoRA method on the Open-Platypus dataset. This fine-tuning aims to enhance its performance across a range of general language tasks.
Key Capabilities & Performance
Evaluated on the Hugging Face Open LLM Leaderboard, this model achieves an overall average score of 47.33. Specific benchmark results include:
- ARC (25-shot): 58.87
- HellaSwag (10-shot): 82.14
- MMLU (5-shot): 54.98
- TruthfulQA (0-shot): 42.84
- Winogrande (5-shot): 77.11
- GSM8K (5-shot): 9.4
- DROP (3-shot): 5.99
These scores indicate a solid foundation in common sense reasoning, reading comprehension, and general knowledge tasks, with a particular strength in HellaSwag.
When to Use This Model
This model is a good candidate for use cases requiring:
- General text generation and understanding: Its fine-tuning on Open-Platypus suggests broad applicability.
- Tasks benefiting from strong common sense reasoning: Indicated by its HellaSwag and Winogrande scores.
- Exploration of Llama 2-based models with QLoRA fine-tuning: Offers a specific implementation for evaluation and deployment.