On Friday, Facebook co-founder Mark Zuckerberg announced the upcoming release of Meta Platforms to researchers of a large new language model called LLaMA (Large Language Model Meta AI). The model developed by the Meta Fundamental AI Research (FAIR) team is designed to help scientists and engineers study AI applications and functions, such as answering questions and summarizing documents.
The LLaMA release comes as tech companies race to promote advances in AI techniques and integrate the technology into their commercial products. As noted by CNBC, the Meta version is different from competitor models as it will come in a variety of sizes, ranging from 7 billion parameters to 65 billion parameters. In addition, Zuckerberg said his company’s new LLM technology – which could eventually solve math problems and conduct scientific research – will be available to the scientific community, with Meta now accepting requests for access. This is a change from Google’s underlying LaMDA and ChatGPT models, which are not publicly available.
Reuters points out that Meta is joining an increasingly intense race to dominate AI technology, which began in earnest in late 2022 with ChatGPT OpenAI. As for Meta, the launch of LLaMA is also an expression of its commitment to open science – hence the decision to publicly release a state-of-the-art basic big language model, along with allowing open resource researchers to advance their work. Meta believes that unlike more finely tuned models designed for specific purposes, their models will prove to be versatile and have many uses.
Another way LLaMA is different, according to Meta: it requires “significantly less” computing power than previous offerings and is trained in 20 languages, focusing on those based on the Latin and Cyrillic alphabets. With 13 billion parameters, LLaMA should surpass GPT-3, the model on which ChatGPT is built. Meta also attributed LLaMA’s performance to “cleaner” data and “architectural improvements” in the model that improved training stability.
To preserve the integrity of the model and prevent abuse, Meta will make it available under a non-commercial license focused on research use cases. Academic researchers, government, civil society, academic institutions and industrial research laboratories will be given model access on a case-by-case basis.
Meta’s launch of LLaMA could mark a significant advance in AI language models. The social media giant’s commitment to open science and allowing researchers to study under a non-commercial license will limit abuses of the model.
LLaMA’s versatility and problem-solving potential can provide insight into the significant potential benefits of AI for billions of people at scale.