Facebook's Powerful Large Language Model Leaks Online (vice.com) 11
Facebook's large language model, which is usually only available to approved researchers, government officials, or members of civil society, has now leaked online for anyone to download. From a report: The leaked language model was shared on 4chan, where a member uploaded a torrent file for Facebook's tool, known as LLaMa (Large Language Model Meta AI), last week. This marks the first time a major tech firm's proprietary AI model has leaked to the public. To date, firms like Google, Microsoft, and OpenAI have kept their newest models private, only accessible via consumer interfaces or an API, ostensibly to control instances of misuse. 4chan members claim to be running LLaMa on their own machines, but the exact implications of this leak are not yet clear.
In a statement to Motherboard, Meta did not deny the LLaMa leak, and stood by its approach of sharing the models among researchers. "It's Meta's goal to share state-of-the-art AI models with members of the research community to help us evaluate and improve those models. LLaMA was shared for research purposes, consistent with how we have shared previous large language models. While the model is not accessible to all, and some have tried to circumvent the approval process, we believe the current release strategy allows us to balance responsibility and openness," a Meta spokesperson wrote in an email.
In a statement to Motherboard, Meta did not deny the LLaMa leak, and stood by its approach of sharing the models among researchers. "It's Meta's goal to share state-of-the-art AI models with members of the research community to help us evaluate and improve those models. LLaMA was shared for research purposes, consistent with how we have shared previous large language models. While the model is not accessible to all, and some have tried to circumvent the approval process, we believe the current release strategy allows us to balance responsibility and openness," a Meta spokesperson wrote in an email.
Leak or PR stunt? (Score:5, Interesting)
Meta is so concerned with this "leak" that there's a pull request with the magnet link sitting here for days:
https://github.com/facebookresearch/llama/pull/73 [github.com]
Zooming in...
https://github.com/facebookresearch/llama/pull/73/commits/56de950af8a48c7cae221581e2e3e2c342b2ad82 [github.com]
Re: (Score:2)
They plan to lay off the team and therefore made the code public, maybe someone will continue developing it for free :)
Re: (Score:1)
Thanks for doing the lord's work.
Unfortunately, there are no seeders.
Re: (Score:2)
There are now ...
Hardly a "leak" (Score:5, Interesting)
The people who will now avoid the approval process are the same people who would have lied through it anyways. But the point of this is they won't be using it under Facebook's approval, which gives them some legitimate cover when the model is used harmfully.
And all the people who will say that's just a phony way for Facebook to CYA can go pound sand. Facebook has no real commercial interest in allowing anybody to have a copy of the model at all. I wish they would all be released openly. It's the parties who use them for harm are the culpable ones.
DIY DGX (Score:2)
I can't wait for the time when I can afford my own DGX's in my closet so I can load the model. I mean what's couple of hundred GBs of VRAM between friends.
Re: (Score:2)
Civil Society (Score:1)
So, everyone in any 1st-world country, except the USA?