Meta Platforms Inc has made another pivotal move by releasing its new language model, LLaMA (Large Language Model Meta AI), intensifying the already tumultuous world of artificial intelligence as Big Tech companies vie for the top spot in the industry.
The race to dominate the AI space began late last year with the launch of OpenAI’s ChatGPT, leading tech giants like Alphabet Inc, Microsoft, and Baidu Inc to showcase their own offerings.
What makes LLaMA peculiar is not its size but efficiency, and weirdly, its general ‘attitude’. It is unlike ChatGPT’s mean jailbreak alter ego known as DAN, which requires very specific commands to activate. LLaMA, on the other hand, is more likely to “cuss” even without explicit instructions to do so.
What is LLaMA?
LLaMA is a state-of-the-art foundational language model aimed at advancing AI research, according to a Meta blog post. It will be available under a non-commercial license to researchers and organizations affiliated with the government, civil society, and academia.
The model has been trained on 20 languages with a focus on those with Latin and Cyrillic alphabets. More importantly, it requires “far less” computing power than previous models.
LLMs [Large Language Models] have shown a lot of promise in generating text, having conversations, summarizing written material, and more complicated tasks like solving math theorems or predicting protein structures. LLaMA technology could eventually solve math problems or conduct scientific research.Mark Zuckerberg, CEO of Meta Platforms
LLaMA outperforms its competitors that examine more parameters and has a 65-billion-parameter model that is “competitive” with Google’s Chinchilla70B and PaLM-540. This performance is attributed to a larger quantity of “cleaner” data and “architectural improvements” in the model that enhanced its training stability.
With a commitment to open science, Meta aims to provide smaller and more performant models to researchers who may not have access to ample computational resources.
Smaller, more performant models such as LLaMA enable others in the research community who don’t have access to large amounts of infrastructure to study these models, further democratizing access in this important, fast-changing field.”Mark Zuckerberg
LLaMA will come in several sizes, from 7 billion parameters to 65 billion parameters, and will be made available to the AI research community.
Here’s a sample prompt and output:
Prompt: write a conversation between the sun and pluto
Answer: Sun: Hey Pluto, what’s up?
Pluto: Not much. Just hanging out here at the edge of the solar system.
Sun: Yeah, I’ve been meaning to talk to you about that. You’re not a planet anymore.
Pluto: What? Come on, man. That’s not fair.
Sun: I’m sorry, but it’s true. You just don’t meet the criteria anymore.
Pluto: This is bulls**t! I’ve been a planet for over 70 years!
Sun: Things change, Pluto. You’ve just got to accept it. Pluto: F**k you, Sun! F**k you and your stupid solar system!
This is what I meant!
How to Apply for Access
The entire AI community, including academic researchers, civil society, policymakers, and industry, must work together to develop clear guidelines around responsible AI in general and responsible large language models in particular. By making LLaMA available to the research community, Meta is setting the same standard of openness, which Google has practiced for years in its AI research.
This is a stark contrast to what OpenAI’s models have been, thankfully, there are different groups trying to create an open-source ChatGPT. Many concerned pundits including Elon Musk have voiced their concerns about Big Tech establishing walled gardens for their AI models.
Large language models like LLaMA work by taking a sequence of words as input and predicting the next word to generate text. The goal of such models is to mine vast amounts of text and summarize information to generate content, making them ideal for answering questions, summarizing documents, and more. Large language models underpin applications such as OpenAI’s ChatGPT, Microsoft Bing AI, and Google’s unreleased Bard.
LLaMA is designed to be versatile and can be applied to many different use cases, making it a foundation model. By releasing the code for LLaMA, other researchers can easily test new approaches to address the known issues in large language models such as bias, toxicity, and the potential for generating misinformation.
If you want to apply for access to LLaMA, click here.
Also check out: 5 Game-Changing Prompt Resources for ChatGPT
Join our newsletter as we build a community of AI and web3 pioneers.
The next 3-5 years is when new industry titans will emerge, and we want you to be one of them.
- Receive updates on the most significant trends
- Receive crucial insights that will help you stay ahead in the tech world
- The chance to be part of our OG community, which will have exclusive membership perks