Model Overview

The dhmeltzer/Llama-2-13b-hf-ds_wiki_1024_full_r_64_alpha_16_merged is a 13 billion parameter language model built upon the Llama 2 architecture. It has been fine-tuned with a focus on general language understanding, evidenced by its performance across a suite of benchmarks. The model supports a context length of 4096 tokens, allowing for processing moderately long inputs.

Key Capabilities & Performance

This model's performance is evaluated on the Open LLM Leaderboard, achieving an average score of 46.33. Specific benchmark results include:

• ARC (25-shot): 58.45
• HellaSwag (10-shot): 81.97
• MMLU (5-shot): 55.02
• TruthfulQA (0-shot): 35.85
• Winogrande (5-shot): 75.69
• GSM8K (5-shot): 10.69
• DROP (3-shot): 6.63

These scores indicate a solid foundation in common sense reasoning, reading comprehension, and general knowledge tasks, while showing areas for improvement in complex mathematical reasoning (GSM8K) and factual accuracy (TruthfulQA).

Good For

• General-purpose text generation and understanding tasks.
• Applications requiring a balance of reasoning and common sense.
• Use cases where a 13B parameter model with a 4096-token context is sufficient.