NousResearch/Meta-Llama-3-8B

Meta Llama 3 8B is an 8 billion parameter instruction-tuned generative text model developed by Meta, optimized for dialogue use cases. It utilizes an optimized transformer architecture with Grouped-Query Attention (GQA) and has a context length of 8192 tokens. Trained on over 15 trillion tokens of publicly available data, this model is designed for commercial and research applications requiring assistant-like chat in English. It demonstrates strong performance across various benchmarks, including MMLU, GPQA, HumanEval, and GSM-8K, often outperforming Llama 2 models of similar or larger sizes.