A new large language model called LLaMA was revealed this week by Facebook’s parent company, Meta, to aid researchers in their work on artificial intelligence.
Researchers and organizations connected to the government, civic society, and academia will have access to LLaMA, or Long Language Model Meta AI, through a non-commercial license.
Massive artificial intelligence (AI) systems known as “big language models” ingest enormous amounts of digital text from news articles, social media posts, and other online sources in order to develop software that can forecast and produce content in response to a prompt or query.
The models can be used to recommend computer programming code, write articles, make tweets, and create chatbot interactions.
According to Meta, LLaMA might outperform rivals that evaluate more factors or parameters than the algorithm does.
It specifically asserted that ChatGPT surpasses GPT-3, a more contemporary model, due to LLaMA’s 13 billion parameters.
According to the business, in order to preserve its integrity and prevent abuse, it is providing the model under a non-commercial license and is currently concentrating on research use cases.
LLaMA has “shown a lot of potential in generating text, conducting conversations, summarizing textual material, and more difficult jobs like solving math theorems or predicting protein structures,” according to Mark Zuckerberg.
With the release of OpenAI’s ChatGPT, backed by Microsoft, in November 2022, the competition for supremacy in the AI technology market formally began. That sparked the release of AI solutions like Bard and ERNIE Bot from internet giants like Google and Baidu.
