The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
thesequence.substack.com
The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
1–10 of 527 posts
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#2I love the human factor, llama was a thing only research would of had access too, of course it leaked and everyone swarms to get it a try and we get some first class support for apple hardware!
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#3Meta before leak: we take safety very seriously and will only give access to responsible researchers
Meta after leak: lol lmfao
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#4This makes is sound as if the Stanford and Berkeley teams also benefited from the leak, whereas I doubt they didn't have official access. So Alpaca/Vicuna/Koala projects would have probably happened anyway. The leak helped with popularity and demand and also somewhat positive PR for Meta, which makes me think they do not mind the leak that much.
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#5The "leak" is being portrayed as something highly subversive done by the darn 4chan hackers.
Before the "leak" Meta was sending the model to pretty much anyone who claimed to be a PhD student or researcher and had a credible college email.
Meta has probably been planning to release the model sooner than later. Let's hope they release it under a true open source license.
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#6> OpenAI published a detailed blog post outlining some of the principles used to ensure safety in their models. The post emphasize in areas such as privacy, factual accuracy
Am I the only one amused by the phrase “factual accuracy”? How many stories have we read like the one where it tries to ghost light the guy that this year is actually last year. “Oh, your phone must be wrong too, because there is no way I could be wrong.” Though, maybe that is what factually accurate means. It is convinced that it is always factually accurate, even though it is not.
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#7> OpenAI published a detailed blog post outlining some of the principles used to ensure safety in their models. The post emphasize in areas such as privacy, factual accuracy Am I the only one amused by the phrase “factual accuracy”? How many stories have we read like the one where it tries to ghost light the guy that this year is actually last year. “Oh, your phone must be wrong too, because there is no way I could b…
> It is convinced that it is always factually accurate, even though it is not.
I don't think that's true. ChatGPT (or any LLM) isn't convinced much of anything. It might present something confidently (which is what most people want) but that's a side-effect of it's programming, not an indication of how good it feels on the answer. If you reply to anything ChatGPT says with "No, you're wrong." it will try to write a new, confident and satisfying answer that responds to your assertion.
LLMs will always be "wrong" because they have no distinction between fiction and fact. Everything it reads is mapped into language, not concept space or an attitude or a worldview.
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#8This makes is sound as if the Stanford and Berkeley teams also benefited from the leak, whereas I doubt they didn't have official access. So Alpaca/Vicuna/Koala projects would have probably happened anyway. The leak helped with popularity and demand and also somewhat positive PR for Meta, which makes me think they do not mind the leak that much.
Meta is actively trying to take down publicly available copies of LLaMA: https://github.com/github/dmca/blob/master/2023/03/2023-03-2...
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#9> OpenAI published a detailed blog post outlining some of the principles used to ensure safety in their models. The post emphasize in areas such as privacy, factual accuracy Am I the only one amused by the phrase “factual accuracy”? How many stories have we read like the one where it tries to ghost light the guy that this year is actually last year. “Oh, your phone must be wrong too, because there is no way I could b…
I’ve been googling trying to figure out what “ghost light” is in this context .. did you get an autocorrect for gas light?
Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT
#10I'm a bit worried the LLaMA leak will make the labs much more cautious about who they distribute models to for future projects, closing down things even more.
I've had tons of fun implementing LLaMA, learning and playing around with variations like Vicuna. I learned a lot and probably wouldn't have got so interested in this space if the leak didn't happen.