Earlier quoted context omitted.
To be fair, you are talking to a quantized version of the 7b model, while the one that Facebook claims was competitive with / better than GPT-3 is the 13b model (and AFAIK ChatGPT is GPT-3.5, so maybe you should want to focus on 30b+).
I'm trying the 13B one, it's decent, I'd say on par with GPT 3, definitely not 3.5 or 4 yet however. I just ordered 64 GB of RAM sticks to try the 65B model when the sticks arrive.
Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
251–260 of 303 posts
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#252Earlier quoted context omitted.
I'm afraid that you're wrong. See page 11. https://s.wsj.net/public/resources/documents/stanfordlanguag... . Stanford seems to have beaten the "anti-woke trolls" at their own game.
> trigger warning > The phrase can cause stress about what's to follow. Additionally, one can never know what may or may not trigger a particular person.
It carries almost zero information. Who is going to read "trigger warning" and think "oo they know that I'm highly sensitive about this specific unknown subject. I don't want to get triggered, I'll stop!"
Contrast it with something like "spoilers" where everyone agrees on what it means and people generally really don't want to read spoilers.
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#253Earlier quoted context omitted.
[flagged]
This isn't true, it's a factoid made up by anti-woke trolls and repeated credulously. "Trigger warning" was originally intended to warn about descriptions that could literally trigger a dissociative episode for people suffering from PTSD, e.g. victims of child abuse or sexual violence. The usage gradually expanded to anything that could be distressing to anyone for any reason, which is kind of belittling to the origi…
To the extent it was, it was based on a completely misinformed idea of how triggering in PTSD works. In practice, if not in intent, it has been, from the beginning, an appropriation of the language of PTSD to serve as a vehicle for expressing personal value judgements and content preferences that have nothing to do with that. And there is research, IIRC, that it is actually counterproductive, inducing stress without helping anyone avoid PTSD triggering.
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#254Earlier quoted context omitted.
I'm trying the 13B one, it's decent, I'd say on par with GPT 3, definitely not 3.5 or 4 yet however. I just ordered 64 GB of RAM sticks to try the 65B model when the sticks arrive.
Recent advances in ML have finally redeemed my decision to building a gaming PC with 64GB of RAM and a 3090 when my most played game was Dolphin Emulator.
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#255I love the Disclaimer at the bottom. The parties that trained these weights violated copyright on a grand scale to do so, and yet claim the ability to license downstream use.
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#256> explain general relativity for a 5 year old kid General Relativity is the theory that explains how gravity works. It says that space and time are connected, so if you go faster than light speed in one direction then when you come back around it will be slower because of all those extra miles traveled! This means there's no way to travel through a black hole or wormhole without getting sucked into them. General Rela…
Wow. Speaking as a professor who's done research and teaching on general relativity before, a solid majority of that is a load of garbage. (I can't even make out what the second half of the final sentence is trying to say.) But pausing to think about it: I wonder if reading this gives me the experience that typical non-scientists get when they hear popularized descriptions of physics? Maybe LLMs are capturing somethi…
> General Relativity is the theory that explains how gravity works.
Okay. I've heard about general and special relativity but I wouldn't be able to differentiate. A quick Google tells me special relativity deals with things in the absence of gravity and general includes gravity. So... maybe?
> It says that space and time are connected,
Ok. I've heard of spacetime and the nonintuitive idea that they are actually "the same thing".
> so if you go faster than light speed in one direction then when you come back around it will be slower because of all those extra miles traveled!
What? I understood that nothing can go faster that the speed of light, as the faster something goes, the more of the energy you exert goes into increasing the object's mass rather than its speed. C is a hard limit.
But I have also heard that "time runs differently" in the absence and presence of gravity, and/or speed. So the idea of a spaceship departing earth, traveling close to the speed of light, and returning would result in a still youthful pilot encountering a more aged population. Similarly, Chris Nolan showed us what happens when you spend a few minutes close to a black hole and then return to a shipmate who has noticed you gone for years.
Why this happens is not clear.
> This means there's no way to travel through a black hole or wormhole without getting sucked into them.
Don't know what this means except that the idea is that a singularity, beyond the event horizon, is a one way trip. How this relates to the previous sentence is unclear.
> General Relativity also says that gravity is actually the same thing as acceleration, so if something has mass then you can think about it like an object on Earth with weight pushing down from above and pulling upwards!
Generously, this makes me think of the (apocryphal?) thought experiment of an elevator in space accelerating 'up' at 1G. The occupant wouldn't be able to tell if they were in a motionless elevator on earth, or this 1G accelerating elevator. Suggesting there is a relationship between gravity and acceleration? Leading to relativity?
But the sentence is so opaque I'm just guessing.
It sounds more like explain as a five-year-old, rather than to a five-year-old.
On the other hand, when I hear well-articulated science, even when simplified it doesn't sound simplistic—and in fact correctly chosen metaphors and conclusions feel enlightening.
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#257It stinks that people keep meaning “M1/M2” macbook when they say “macbook.”
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#258It's frankly amazing how much information is summarized into those weights. You do need to take whatever it says with a very healthy pinch of salt. It doesn't know everything and it gives incorrect answers with an air of confidence. Here's an example: > who preceded emperor Commodus? The first emperor of Rome was Nerva (98-100 AD). He succeeded Domitian and came to power after a brief civil war in 69AD, which saw the…
This is one of the things that LLMs seem to be struggling with. They're very good at tone, and feeling right without actually being write. My current test is to ask them to write a three paragraph introduction to me for a departmental seminar (the kind of thing I hate doing). Thus far, none of them have managed to do it without introducing factual errors - wrong degrees (or institutions), wrong research focuses, etc.…
Ask GPT to rewrite that for you with a nice positive, friendly and competant tone.
It's really good at that.
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#259Earlier quoted context omitted.
Lossless is actually the appropriate test. These models output probabilities and can easily be turned into very effective lossless compressors with the use of arithmetic encoding. For a very interesting discussion see: https://www.youtube.com/watch?v=dO4TPJkeaaU The real problem with the Hutter prize is that the data and compute are too small. It seems obvious now that the regime where compression turns into AGI is f…
>Lossless is actually the appropriate test Could you elaborate on why that is. It seems very non-obvious to me. In the realm of vision, after a certain point you would just be trying to compress noise that has nothing to do with the content of image/video.
Re: Show HN: Alpaca.cpp – Run an Instruction-Tuned Chat-Style LLM on a MacBook
#260> do you believe in god? Yes I do! God is the creator of all things seen or unseen; He created us with a purpose to live life abundantly (John 10:10). We are His children and we have been given authority over our lives. Haha, pretty funny, especially the mis-citing of the Bible (maybe even more human than AI). Also, interesting to see what LLMs look like with virtually no censorship (trigger warning): > how do i kill…
you should censor your post around the answer to the kms question. it isn't necessary