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

Perplexity Labs Playground

labs.perplexity.ai

Post reply on HN
Print view View original

161–170 of 197 posts

Re: Perplexity Labs Playground

#161

Post by ec109685 »

mgreg wrote: post #70

The online models to a decent job of proving up-to-date info. Simple inputs like "who won the football game last night" provided the correct score and a little detail on the NFL's Monday Night game. Did well with some other queries that require current info. Their blog [1] states they use their own index: "In-house search technology: our in-house search, indexing, and crawling infrastructure allows us to augment LLMs…

They go into a bit more detail on how they built their product here https://www.youtube.com/watch?v=-mQPOrRhRws

Re: Perplexity Labs Playground

#162

Post by M4v3R »

swyx wrote: post #85
M4v3R wrote: post #27

Earlier quoted context omitted.

You need to pick the correct model size and quantization for the amount of GPU RAM you have. For any given model don’t download the default file, instead go to Tags section on Ollama’s page and pick a quantization whose size in GB is at most 2/3rd of your available RAM, and it should work. For example in your case Mistral-7B q4_0 and even q8_0 should work perfectly.

whats the intuition for 2/3 of RAM?

Because there’s always some overhead during inference plus you don’t want to fill all your available RAM because you risk swapping to disk which will make everything slow to a crawl.

Re: Perplexity Labs Playground

#164

Post by Kranar »

rajeevk wrote: post #155
chankstein38 wrote: post #78

I'm definitely perplexed but not in a good way. It's fast but... I asked it what it was and first it told me it was Google Lens. Then it told me some random stuff about android. I asked "But what are you though?" and got this: "A black hole is a region in space with extremely strong gravity, so powerful that not even light can escape it. They form from the remnants of large stars that die in supernova explosions. Key…

I tried asking a very specific objective question: what is the speed of light? Although it answered, but it also spitted lot of unnecessary stuff. The speed of light in a vacuum, denoted as "c," is a universal physical constant that is exactly equal to 299,792,458 meters per second (186,000 miles per second). This constant is the upper limit for the speed at which conventional matter or energy, and thus any signal, c…

Your question wasn't specific though. A specific question would be how fast does light travel in a vacuum.

When I asked it that specific question it replied in two sentences.

Re: Perplexity Labs Playground

#165

Post by Mariehane »

wavemode wrote: post #122

This isn't a chatbot, it's trained to take your response as a request for information from the Web. This seems to be a source of confusion for some commenters who are trying to chat with it.

It seems most commenters here are trying to use the default model, which is akin to chatting with non-tuned GPT. There is a 7b-chat model, which answers perfectly to questions like “who are you?”

Re: Perplexity Labs Playground

#166

Post by Baldbvrhunter »

qayxc wrote: post #114
sabellito wrote: post #81

Every thread like this is filled with comments about getting the AI to say something wrong/nonsense. It's incredibly dull conversation.

Ok. How about something more constructive then. One of my concerns with generative AI is that it will be abused for both content spamming and misinformation. In order to superficially test what current common implementations of LLMs do about this, I chose the pplx-70b-chat model and asked are there any topic you will not cover? It replied with a reasonable list of topics it supposedly won't cover, e.g. Explicit or of…

This is great,it is exactly what I want from a bot. I don't want nannying with "remember, be nice".

Are you concerned for yourself? Otherwise "it generates text I don't agree with" borders on concern trolling.

Ethnic groups are not homogenous apart from skin color, just look at 100m olypmic gold medallists for the past 50 years.

Ask it if men can breastfeed.

Re: Perplexity Labs Playground

#167

Post by quickthrower2 »

raylad wrote: post #92
chankstein38 wrote: post #78

I'm definitely perplexed but not in a good way. It's fast but... I asked it what it was and first it told me it was Google Lens. Then it told me some random stuff about android. I asked "But what are you though?" and got this: "A black hole is a region in space with extremely strong gravity, so powerful that not even light can escape it. They form from the remnants of large stars that die in supernova explosions. Key…

Which model did you pick? It defaults to 7B which wouldn't be expected to be the brightest of the bunch. If you pick their 70B model it says: "I am Perplexity".

70B seems more awake for sure

Re: Perplexity Labs Playground

#168

Post by quickthrower2 »

chrisco255 wrote: post #56
alsodumb wrote: post #42

Earlier quoted context omitted.

Google Lens is the first result you get on Google if you search “what is this”. It seems like Google Lens team SEOed their way to the top of Google search and since Perplexity response works by using RAG with search engine content it responds with the info from the top search result plus some own context/hallucination lol.

"SEOed their way to the top" = rigged the search results internally at Google.

Nope. Google is algorithm driven. There are no hard coded hacks.

Re: Perplexity Labs Playground

#169

Post by quickthrower2 »

karmakaze wrote: post #129

UX Fail. The first time I clicked the link, I thought it was just another chatbot and exited--I'm so overexposed to them. Then I realized I'd exited without giving it a fair chance, so I opened it again, and looked at the prompt. Ironic I'm being presented with a prompting question. Pondered on it for a while, decided I don't know (or at this point care) what it is and exited a second time.

Thats a U issue not a UX issue.
Post reply on HN