The race to establish dominance in the AI technology space began in late 2022 with the launch of OpenAI’s ChatGPT, which received backing from Microsoft. The launch spurred tech giants such as Alphabet Inc (GOOGL.O) and China’s Baidu Inc (9888.HK) to announce their own AI language models.
One of the key features of LLaMA is its ability to operate in multiple languages, making it highly adaptable to various contexts and scenarios. In addition, LLaMA incorporates a novel training algorithm that can handle large amounts of data in a highly parallelized manner. This approach allows for faster and more efficient training of language models, which can be applied to a wide range of natural language processing tasks.
The researchers conducted experiments using the LLaMA model on several benchmark datasets, including the popular GLUE and SuperGLUE benchmarks and achieved state-of-the-art results on several tasks. The results showed that LLaMA is highly effective at handling complex language tasks while maintaining a relatively small model size and fast inference time.
According to the company, LLaMA can outperform competitors that examine more parameters, including Microsoft-backed OpenAI’s ChatGPT. The 13 billion parameter version of LLaMA is said to outperform GPT-3, a recent predecessor to the model on which ChatGPT is built.
This new development marks another step in the ongoing battle to dominate the AI technology space, which has seen companies like Alphabet Inc and China’s Baidu Inc vying for dominance with their language models. With its impressive performance and large parameter count, LLaMA may give Meta an edge in this competition. However, it remains to be seen how other companies will respond to this latest development.
Original article: https://news.coincu.com/170087-meta-llama-outperforms-including-chatgpt/
