Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

thesequence.substack.com

Post reply on HN
Print view View original

241–250 of 527 posts

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#241

Post by jonplackett »

drumhead wrote: post #228
jonplackett wrote: post #162

Earlier quoted context omitted.

The difference between 3.5 and 4 is gigantic even in my fairly limited experience. I gave them both some common sense tests and this one stuck out to me. Q: A glass door has ‘push’ written on it in mirror writing. Should you push or pull it GPT-3.5: If the word "push" is written in mirror writing on a glass door, you should push the door to open it GPT-4: Since the word "push" is written in mirror writing, it suggest…

I tried the same question with Bard, and it got it right but it also gave me the following: Here are some other things to keep in mind when opening a glass door: Use your hands to open the door. Do not use your feet or other body parts. Be careful not to break the glass. Glass doors can be very fragile, so it is important to be gentle when opening them. If you are unsure how to open a glass door, ask for help from so…

Poor Bard. I felt quite sorry for it while reading that.

It’s funny how with these human-like systems you get a gut feeling about their intelligence before you have any hard evidence.

My 3 year old worked out Siri is dumb compared to Alexa

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#242

Post by famouswaffles »

PoignardAzur wrote: post #190
jonplackett wrote: post #162

Earlier quoted context omitted.

The difference between 3.5 and 4 is gigantic even in my fairly limited experience. I gave them both some common sense tests and this one stuck out to me. Q: A glass door has ‘push’ written on it in mirror writing. Should you push or pull it GPT-3.5: If the word "push" is written in mirror writing on a glass door, you should push the door to open it GPT-4: Since the word "push" is written in mirror writing, it suggest…

Wait, how does GPT-4 even... Does it benefit from its visual attention, or is it a case of "the question wasn't in GPT-3's training set but it was in GPT-4's"?

It's pretty much looking like anything can be extracted from language. Some harder than others for sure but with enough scale it does look like eventually everything falls. Text only GPT-4 has a pretty solid understanding of space that 3.5 definitely lacks. You can see more thorough experiments in the microsoft agi paper where they test it's ability to track the visual space of a maze.

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#243

Post by FLT8 »

jonplackett wrote: post #236
FLT8 wrote: post #211

Earlier quoted context omitted.

This is what I get out of Vicuna-13b: Human: A glass door has 'push' written on it in mirror writing. To open the door should you 'push' or 'pull' it? Assistant: Since the word "push" is written in mirror writing on the glass door, you should actually "pull" the door open instead of "push" it. Mirror writing is a writing method where the characters are reversed, so when you see the word "push" written in mirror writi…

The difference for me with GPT-4 is its ‘understanding’ of the scene and its explanation of WHY you should push or pull. It talks an out a door with people approaching from different directions. It has some idea of what those people would be thinking. That seems different to just ‘mirror writing means do the opposite’.

Maybe, but consider the post above where GPT4 gets confused about the blind man on the other side of the door, while Vicuna-13b seems to figure it out. I accept that GPT4 gave a better answer in this case, but its level of understanding about the scene under different scenarios still seems limited.

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#244

Post by jonplackett »

tel wrote: post #206
steveBK123 wrote: post #202

Earlier quoted context omitted.

The evolution of answers from version to version makes it clear there are insane amounts of manual fine tunings happening. I think this is largely overlooked by the "its learning" crowd.

This is an interesting argument as it's easy to apply it nearly universally to any example of learning. What sort of evidence would convince you that it is learning?

I think this is a great question we should all think about for ourselves in advance - what does it have to do to convince you it’s actually intelligent.

Because once it does that thing without you having expressly decided that is the goal, it’s very tempting to just move the goal a liiiitle bit further away

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#245

Post by int_19h »

sbierwagen wrote: post #238
barbariangrunge wrote: post #168

Earlier quoted context omitted.

I thought it needed 64gb of vram. 64gb of ram is easy to obtain

5950x is a CPU model. Integer-quantized models are generally run with CPU inference. For the larger models the problem then becomes generation time per token.

Quantized models are used aplenty with GPUs as well - 4-bit quantization is the only way you can squeeze llama-30b into 24Gb of VRAM (i.e. RTX 3090 or 4090).

In fact, I would say that, at this point, most people running LLaMA locally are likely using 4-bit quantization regardless of model size and hardware, just to get the most out of the latter.

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#247

Post by int_19h »

jhbadger wrote: post #37
anonzzzies wrote: post #31

Earlier quoted context omitted.

It is vastly better than anything else so far though. The rest will catch up but openai is not sleeping and they are well funded.

I thought that was the case before trying Vicuna. I agree that LLaMA and Alpaca are inferior to ChatGPT but I'm really not sure Vicuna is. It even (unfortunately) copies some of ChatGPT's quirks, like getting prudish when asking it to write a love scene ("It would not be appropriate for me to write...")

It's definitely inferior once you start giving it non-trivial tasks. the same goes for all instruction- and chat-tuned forks of LLaMA. Maybe that'll be different on larger model sizes; we haven't seen vicuna-30b or 65b yet. But if it really is that easy, I'd expect there to be a lot more commercial offerings on par with ChatGPT already.

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#248

Post by klyrs »

JumpCrisscross wrote: post #136
klyrs wrote: post #119

Earlier quoted context omitted.

As far as I can tell, you can claim credit for the output of a tool which produces something at your direction. If you write an algorithm to generate patents, and you execute that algorithm and submit its output to USPTO under your own name, no problem. If you gave credit to the algorithm, their policy is to deny the claim.

Can OpenAI claim copyright on GPT’s outputs?

I am not a lawyer, mind, but if they're trying to claim sole copyright the output of the machine's collaboration with its users, that sounds awfully dubious. Knowing lawyers, and having some familiarity with copyright law, the only thing i can confidently say is that your question will take several very expensive years to answer.

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#249

Post by int_19h »

jhbadger wrote: post #192
sp332 wrote: post #175

Earlier quoted context omitted.

This is incorrect. According to the official https://github.com/lm-sys/FastChat#vicuna-weights you need the original Llama weights before applying the Vicuna diff.

Seriously, you can download the Vicuna model and run it locally with llama.cpp. I've done it!

It doesn't matter if you merge the LoRA, the resulting weights are still a derived work - assuming, that is, that weights are copyrightable in the first place (which is still a big if).

Re: The LLama Effect: Leak Sparked a Series of Open Source Alternatives to ChatGPT

#250

Post by int_19h »

yieldcrv wrote: post #121
greyface- wrote: post #113

Earlier quoted context omitted.

They have tacitly endorsed the leak. https://github.com/facebookresearch/llama/pull/73#issuecomme...

Only because publicly visible actions are worse for them People have gotten DMCA takedown requests from them over Llama repositories

Are they going after copies of LLaMA, or after LLaMA derivatives specifically?
Post reply on HN