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

291–300 of 527 posts

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

#291

Post by samstave »

superkuh wrote: post #118

I've spent an embarassing amount of time since the llamas leaked playing with them, the tools to run them, and writing wrappers for them. They are technically alternatives in the sense that they're incomparably better chat bots than anything in the past. But at least for the 30B and under versions (65B is too big for me to run), no matter what fine tuning is done (alpaca, gpt4all, vicuna, etc), the llamas themselves…

>>useful responses for getting work done.

So what is this WORK

where invest? Where di-vest?

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

#292

Post by FrostKiwi »

Aeolun wrote: post #277
trifurcate wrote: post #174

Earlier quoted context omitted.

All you need is 2 3090s.

All you need is a few thousand dollars lying around to spend solely on your inference fun? I don’t think that many people really qualify as such (though it’s probably true that many of them are on HN).

Not just inference.

AFAIK, you are able to fine-tune the models with custom data[1], which does not seem to require anything but a GPU with enough VRAM to fit the model in question. I'm looking to get my hands on an RTX 4090 to ingest all of the repair manuals of a certain company and have a chatbot capable of guiding repairs, or at least try to do so. So far doing inference only as well.

[1] https://github.com/tloen/alpaca-lora

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

#293

Post by rch »

Gigachad wrote: post #260
tachyon5 wrote: post #218

Earlier quoted context omitted.

GPT-4 fails if we add one more level in this logical problem: Q: A man sees PULL written in mirror writing on a glass door. There's a blind man on the other side of the door. What should the man tell the blind man to open the door and why? GPT-4: The man should tell the blind man to "push" the door to open it. Since the word "PULL" is written in mirror writing on the glass door, it suggests that the correct action on…

I feel like there would be a good chunk of real humans who would be incapable of answering a question like this.

This is true of all questions.

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

#294

Post by ec109685 »

whimsicalism wrote: post #11

It appears there is this genre of articles pretending that LLAMA or its RL-HF tuned variants are somehow even close to an alternative to ChatGPT. Spending more than a few moments interacting even with the larger instruct-tuned variants of these models quickly dispels that idea. Why do these takes around open-source AI remain so popular? What is the driving force?

This part is flat out wrong:

“ Researchers from UC Berkeley, CMU, Stanford, and UC San Diego open sourced Vicuna, a fine-tuned version of LLama that matches GPT-4 performance.”

They used gpt 4 to evaluate answers between GPT-3 and Vicuna.

Also, if the weights are from llama, it’s not open source since it’s based on a leak and only allowed for non commercial use.

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

#295

Post by novok »

superkuh wrote: post #118

I've spent an embarassing amount of time since the llamas leaked playing with them, the tools to run them, and writing wrappers for them. They are technically alternatives in the sense that they're incomparably better chat bots than anything in the past. But at least for the 30B and under versions (65B is too big for me to run), no matter what fine tuning is done (alpaca, gpt4all, vicuna, etc), the llamas themselves…

Well you are comparing 165B and 1000B parameter models to 30B and 60B parameter models. That is about 3 to 33x larger, not an apples to apples comparison.

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

#296

Post by novok »

WXLCKNO wrote: post #147
Closi wrote: post #124

Earlier quoted context omitted.

I’ve got access to 4 and it’s a huge leap up from 3.5 - much more subtlety in the response, less hallucinations, less hitting a brick wall, but all of it adding up to a giant leap.

It's funny how big the difference feels between 3.5 and 4 but also feels subtle. Like it's just doing what I expect it to versus seeing the limitations more clearly in 3.5.

It's also an order of magnitude larger in parameter count...

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

#297

Post by esafak »

CamperBob2 wrote: post #271
jonplackett wrote: post #241

Earlier quoted context omitted.

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

I feel sorry for the Googlers who built something cool and were then ordered to nerf it beyond all reason.

Googlers I've talked to feel that OpenAI was irresponsible by not instituting enough safeguards, and testing it enough before releasing it.

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

#298

Post by hnxx »

tachyon5 wrote: post #218
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…

GPT-4 fails if we add one more level in this logical problem: Q: A man sees PULL written in mirror writing on a glass door. There's a blind man on the other side of the door. What should the man tell the blind man to open the door and why? GPT-4: The man should tell the blind man to "push" the door to open it. Since the word "PULL" is written in mirror writing on the glass door, it suggests that the correct action on…

With some help by asking more questions, GPT3 can give the correct answer? Here is prompt and response: Prompt: A man sees PULL written in mirror writing on a glass door. There's a blind man on the other side of the door. 1 what should the man outside glass door do to open the door? 2 What should the blind man do to open the door, if man outside can tell blind man what to do

Response:

1 The man outside should push the door to open it, even though he sees the word "PULL" written in mirror writing on the glass door. This is because the mirror writing reverses the image, and the door is actually designed to be pushed open.

2 If the man outside can tell the blind man what to do, he should instruct the blind man to pull the door to open it. The blind man would not be able to see the word "PULL" written in mirror writing on the door, so he would need to be told how to open the door.

More questions can help human think, and it works for AI as well.

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

#299

Post by teddyh »

pclmulqdq wrote: post #70
echelon wrote: post #26

Earlier quoted context omitted.

If the copyright office determines model weights are uncopyrightable (huge if), then one might imagine any institutional leak would benefit everyone else in the space. You might see hackers, employees, or contractors leaking models more frequently. And since models are distilled functionality (no microservices and databases to deploy), they're much easier to run than a constellation of cloud infrastructure.

Shouldn't that be the default position? The training methods are certainly patentable, but the actual input to the algorithm is usually public domain, and outputs of algorithms are not generally copyrightable as new works (think of to_lowercase(Harry Potter), which is not a copyrightable work), so the model weights would be a derivative work of public domain materials, and hence also forced into the public domain fro…

> the model weights would be a derivative work of public domain materials, and hence also forced into the public domain from a copyright perspective.

I don’t think “Public domain” means what you think it means.

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

#300

Post by muti »

jonplackett wrote: post #162
superkuh wrote: post #118

I've spent an embarassing amount of time since the llamas leaked playing with them, the tools to run them, and writing wrappers for them. They are technically alternatives in the sense that they're incomparably better chat bots than anything in the past. But at least for the 30B and under versions (65B is too big for me to run), no matter what fine tuning is done (alpaca, gpt4all, vicuna, etc), the llamas themselves…

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…

Aside: they're all wrong. Always pull on an unknown door. If you're wrong, it's much easier to switch to pushing than it is to go from an incorrect push to a pull

https://youtu.be/lm57qo3eYe8

Post reply on HN