Archive by category "Chat GPT"

💭💻What does Jordan Peterson think about AI? The answer might surprise you!

Is Jordan Peterson right about AI? Learn Jordon Peterson think about AI now! 🤖🤔

TL;DR Summary: In a YouTube video, Jordan Peterson think about AI and expresses his concerns about artificial intelligence (AI) and what it means for humanity. He argues that the bigger the technological toy, the smarter we must be to use it. He describes how virtualization enables psychopathy and how anonymous trolls on Twitter are psychopathic, Machiavellian, narcissistic, and sadistic.

Jordon Peterson notes that the pace of technological advancement is so fast that we could see an AI giant walk around the earth in the next two to three years. He believes that we are primitive children on the ethical front with the tools of gods, and we need to take on an ethical responsibility commensurate with our technological power. Otherwise, we could make this world quite a heavenly place, but there are a lot more pathways to hell than there are to heaven.

Jordan Peterson’s Thoughts on Artificial Intelligence

We must use technology responsibly to prevent the potential for corruption and misuse of power.

00:00 🤔 We must use tech wisely to avoid creating powerful tools that could lead to giants walking the earth again.
- We must be smarter than our technology to use it responsibly, as we are already creating powerful tools that could potentially lead to giants walking the earth again.
00:58 🤔 Anonymous trolls on Twitter have dark personalities.
- Anonymous trolls on Twitter have been studied and found to have the dark tetrad of personality traits, including Machiavellianism, narcissism, and sadism.
02:28 🤯 Virtualization amplifies psychopathy and narcissistic Machiavellianism.
- Virtualization enables psychopathy, which is tearing apart our culture and amplifying the pathology of narcissistic Machiavellian people.
03:55 🤔 We must reevaluate Good and Evil in light of our technological power and take on an ethical responsibility.
- Alchemy’s promise of matter led to the emergence of science, which has made us technologically advanced, but also carries an ethic from the Judeo-Christian tradition.
- We must take on an ethical responsibility commensurate with our technological power, requiring a revaluation of Good and Evil.
05:37 🤔 We must use tech responsibly or face dire consequences.
- We are at a crossroads and must be careful with our technological power or risk a nightmare beyond contemplation.
06:39 🚫 Individuals must take responsibility to prevent corruption.
- Individuals must take responsibility for their actions to prevent the corruption of the world.
07:53 🤯 Everyone is responsible for their own and others’ actions.
- Everyone is responsible for their own actions and the actions of others, a realization that can be overwhelming and even lead to psychosis.
08:40 🤔 Choose wisely: our words create our reality and determine our fate.
- People must choose between accepting responsibility for their actions or believing their life is ultimately meaningless.
- Every word we say creates reality and tilts the world towards heaven or hell, so be careful with what we say.

Here are the full transcripts of Jordon Peterson’s view of AI

(5) Jordan Peterson’s Thoughts on Artificial Intelligence – YouTube

Transcript:
(00:00) can you comment on your thoughts about artificial intelligence and what it means for Humanity the bigger the toy the smarter you better be to use it so that’s and that’s worth thinking about because we’re coming up with some pretty damn big toys and one of my friend Jonathan pajo said again he said recently he’s I told you about him this deep deep religious thinker he’s something he said we’re going to see giants walk the earth again and you think well what does that mean it’s like well

(00:47) we’re already Androids you know already when you’re on Twitter you’re an Android well why well you know there are lots of trolls on Twitter sometimes I’m one of them by the way but I’m not an anonymous troll so I take my lumps let’s say well I started calling the trolls troll demons and uh I had a reason for that and the reason is you know maybe you’re some resentful character who’s a coward and and your Machiavellian and narcissistic and Psychopathic and sadistic because that’s actually the

(01:26) Hallmarks of anonymous trolls and that’s been studied very well by people who study the so-called dark tetrad of personality traits I interviewed this guy Dale paulus who’s the head researcher among the dark dark tetrad research group and that’ll be coming out right away and it’s clearly the case that the anonymous denigrating trolls are Psychopathic Machiavellian narcissistic sadists and maybe not right to the clinical point but enough so that you can detect it using the proper psychometric measurement instruments and

(02:01) so you think you got this resentful you know underground character who’s carpent away on Twitter trying to do nothing but cause trouble who’s also Anonymous and he’s bitter and he wants to hurt people and he likes to hurt people which is what lulls means right I’m doing it for the lulls means I’m doing it because I positively Delight in the production of human misery and you think if you just met that guy you wouldn’t give him a half a second of your time in fact he wouldn’t even meet him because

(02:29) he’s cowering in his mother’s basement covered with Cheeto dust under the covers you know and he’s just he’s just there’s nothing to him at all except bitterness in spite but now you multiply him 10 million times and you make him electronic and that’s an Android and that’s what happens every time you post something on Twitter and she’s not it’s it’s he’s already a human machine hybrid and he’s a human super intelligent connected machine hybrid and so that means all his

(02:57) pathology is Multiplied immensely by our Tech by our technological might and I’ve thought recently that virtualization enables psychopathy so in normal social situations if you’re a narcissistic Machiavellian and you push it too hard someone will absolutely hit you and that will be the end of it and most of the time it’s cowards who take that route and it doesn’t take much intimidation to stop them just the hint of appropriate just retaliation will keep them silent but online especially if you’re

(03:32) anonymous that’s all Stripped Away and so the Psychopaths can have free reign and part of the reason that our culture is being torn apart at the moment is because virtualization enables psychopathy and so that’s a big problem and so that all pertains to this question of well you know when you have bigger toys you better be a bigger boy and that’s something Carl Jung commented on in detail in his writing he thought that he studied Alchemy you know thousands of years of alchemical writing the Alchemists were motivated by the idea

(04:03) that there was a substance that was material and if we could investigate material and discover that substance we would live forever we wouldn’t be ill and we could transmute base Metals into gold which would be infinitely wealthy and Jung’s notion was that the promise of matter was what impelled Alchemy and out of that grew science and science is motivated by the notion that if we study matter deeply enough we’ll prolong life will increase health and will make ourselves wealthy and of course that’s exactly what

(04:36) happened and we took this technological route there so out of alchemy this weird dreamlike mythology emerged science and just blew up and made us into technological Giants but there was an ethic in alchemy too that was Allied in part complexly with the whole judeo-christian tradition there was an ethic and it didn’t expand at the same rate and so now we’re you know we’re primitive children on the ethical front with the tools of gods and that’s very very dangerous and in fact it might be fatally dangerous and so one of the

(05:12) things that Jung attempted to do in his work was to help people understand that it was necessary for us to take on an ethical responsibility that was commensurate with our technological power and that that would require something like a something like the deepest of all possible religious revaluations right the revaluation of Good and Evil to use the nietzschean phrase as an as of what medication for the death of God is very serious set of ideas and you think we’re powerful now man you wait till the AI giant starts to stroll

(05:46) around the earth and that’s common like you will absolutely see that in your lifetime you’ll probably see it in the next two or three years because that technology is moving so quickly that even those who are on top of it can’t keep up so and I would say we better get our act together because we could make this world quite the Heavenly place but there’s a lot more Pathways to Hell than there are to have it and you know by the time World War II hit we were pretty damn technologically powerful we could

(06:14) take out a lot of the world in a very short period of time but that was nothing compared to what we can do now so we’re we’re really at a Crossroads we could have everything we could dream of if we were careful but we could have a nightmare Beyond contemplation if we’re not and so I would say and I would say this is all on you right because also as we become technologically more powerful every single individual becomes more powerful and that means the corruption of the individual starts to become more and more a world

(06:46) destabilizing Force well obviously right because look how connected you are so how could it be otherwise right because that connection means your power is Multiplied obviously that’s what it means and that means your sins are multiplied really now we already know the people who studied totalitarian states with any degree of depth figured out 40 years ago that a totalitarian state can only Thrive if every single person in the state lies solzhenitzen claim was one man who stops lying can bring down a tyranny which he

(07:19) did by the way now not alone you know Reagan had something to do with that so did so did the Pope at that time and not a variety of other people as well but they were three Cardinal individuals and so you know you look there look out there in the world and you see the Giants marching towards you and you think well what can I do and the answer is bloody well get your act together because here it comes boys and girls and you better be prepared so you know and and also not just for your own sake but for the sake of everyone

(07:53) what did Dostoevsky say everyone is responsible not only for what they do but for what everyone else does and that’s that’s an insane thing to say but Dostoevsky was an insane genius and with heavy emphasis on the genius and it’s very weird to understand that the world might be constituted so that everything that you do matters so much that it also affects everything everyone else does but that is how the world’s constituted and that’s a hell of a burden to take on and a very frightening realization you

(08:24) know it’s enough to make you psychotic psychotic if it comes upon you unawares you know people who become manic they get some sense of their of the potential of their reach but it comes upon them involuntarily and it just destroys them it’s just they’re too close to God that’s a good way of thinking about they’ve flown too close to the Sun Sun’s still there and you need it but you can definitely fly too close to it but you’re still stuck with the responsibility and that’s a very daunting thought you

(08:53) know I’ve often thought if people had to choose between imagine you could choose here’s your choice your life is meaningless fundamentally ultimately you know you’re just a dust spec in in a in a you’re just a grain of sand on an infinite Beach and and what you do doesn’t really matter and that’s very dismal view of the world but on the other hand you don’t have any responsibility you can do whatever you want and so you got to ask yourself when people make that claim is like are they Stark realists are they

(09:25) just trying to evade responsibility and there’s all sorts of reasons to evade responsibility so I would use that hypothesis first and foremost and the alternative to that is an even more terrifying View and that’s associated with the fear of God that’s the beginning of wisdom is every single thing you do matters it echoes and and spreads every word you say every word you say constitutes reality that’s the Divine word every word you say tilts the world towards heaven or hell and that’s the truth so

(10:02) [Applause] so be careful what you say right because you’re bringing the world into being from the realm of unformed possibility and you’re making a decision whether that’s a tilt towards heaven or a tilt towards hell and really that’s what you’re doing so someone’s excited about that hopefully someone who’s aiming upward

Chat GPT AutoHotkey tools: Unlock the full potential of ChatGPT3 with these must-have amazing AHK Tools❗

🚀Unlock the full potential of ChatGPT3 with these must-have AutoHotkey tools!🔥

This video is about utilizing Chat GPT AutoHotkey tools to enhance the experience of using ChatGPT3.

Joe, introduces two scripts that he created: “Chat GPT summary” and “AutoHotkey rules.” The Chat GPT summary script allows users to quickly share the titles of the latest activities in ChatGPT with a recipient of their choice by pressing Ctrl + G. The AutoHotkey rules script is used to set instructions for ChatGPT3 to follow. The speaker also mentions an extension called “Share GPT” that adds a “Share” button, allowing users to share the whole thread with others in a formatted and easy-to-follow manner. The speaker emphasizes that the information in each thread is lost when a new thread is started, so uploading the information into the user’s own environment is the next big step.

🚀Unlock the full potential of ChatGPT3 with these must-have Chat GPT AutoHotkey tools !🔥

And get the summary

Transcript: Chat GPT AutoHotkey tools

(00:00) hey everyone it’s Joe this is here from the automator and today we got some really cool stuff really focused around mostly Auto hockey with chat cheap BT and there’s a couple tools that we wrote and then one extension that comes in really handy using it so why don’t you go ahead and show the the first one here okay so I just created two scripts one of them is the chat upd summary it allows us to it helps you share the titles of the latest things that you have done in chat GPT because we we wanted to kind of like help you share

(00:34) ideas on how you’re using the tool so it is very simple you go to chat GPT you make sure that the bar is visible here on the left what you do is you press Ctrl G and it would automatically read all the titles on the left bar and just create a quick email this email address here can be configured to somebody else so part of of the ideas that you can set any recipient that you want and the idea is to have a quick summary of how you have been using chat GPT on the past few days just to get interesting ideas on

(01:14) how you can actually use it so this is a very quick example of uh using the interface the UIA interface to grab stuff from the screen right so in this script if you actually study the script you learn you can you know possibly learn a little bit about using UIA also about using the Run command with an email address the the one trick is the people have to have like the default client yes that is correct if not um it does get copied to your clipboard so if it doesn’t work for you just go to an email and paste and it’ll have the

(01:48) content right you should have the content in in your clipboard after use so if you do not have a default client installed it doesn’t matter you still have the data there now um the other one that we discussed was the one for creating kind of like little rules for Auto hotkey for chat GPT-3 sorry so we tell chat I thought we tell GPT that autohockey rules as well but here’s the thing so sometimes when when you tell it to write screen ribs there’s a few things that she does a little bit you know she’s

(02:25) getting a little bit creative so what we do is that we created a hot string which of course you can change but this is kind of like a a setup I just tell her that when she’s gonna start she just follows this set of instructions and after I do that every single one of the prompts that I create after that and actually uh yeah she will know that she should follow one of those preferences of course you can change those to anything you like but this is just at least an idea of what you can tell her to go ahead and follow your instructions

(03:01) for every single script she creates within that thread right yes in that thread that’s the important part about chat GPT-3 is it’s it’s really a garbage and garbage out kind of thing if if you don’t set the stage and give it some things to say make sure you do it in this way it’ll take Liberty and do it however it wants to and often it makes up its own things and it doesn’t actually really write the functions so we put in a lot of things that just help kind of instantiate it right to guide it

(03:28) yeah get it ready and this way when you actually use it to create a hockey script it is more likely to give you a working Auto Auction script with functions that actually exist uh it’s not perfect but it does help we also did a little testing that you can keep going and going in this one thread and for the most part it seems to follow along so you can have basically an auto hotkey thread right you can do this once now we put it into a hot string so we can just throw it in whenever we want to or uh to dump it in there but that is correct now

(04:01) uh the last one that I think we is not another hotkey script or anything it is a an extension that I downloaded if I remember correctly was the shared Chat GPT here it is share GPT and that extension adds an additional button here as you can see that says share and what it does is that it grabs that thread the whole thread and allows you to have kind of like a shareable link that you can go ahead and paste it could be on a on a different computer to another person and they could see the whole thread so in case that you

(04:39) found you did something that is really uh interesting like hey how how do I use this function or whatever and you want to share that with somebody else you do not have to copy paste or you know you just hit this share button and you get the whole threat everything that you did up to the point that you finished is going to be in the thread and it is really it keeps the formatting so it’s really easy to follow so that was the other one that we wanted to kind of like share yeah so uh let us know if you found these useful and how you’re using Chat GPT-3

(05:10) if you want to comment below if you learned something here about these Chat GPT AutoHotkey tools please like the video it really helps us out and uh yeah we’re we’re still hitting it hard with chat GPT we’ve got it where we’re uploading files into our own environment that’ll be the next real big aha is where we’ve so when you do this you are in this thread teaching customizing chat gbt it’s learning this is what I want with this thread but once that thread ends it goes away right right as soon as you start a new thread

(05:39) then it is kind of like reset completely so you can go back to that thread don’t get me wrong and they’ll still have that but the point being like it’s it’s kind of a stale thing now we’re uploading into our own environment and uploading files and doing stuff that’ll be our environment that doesn’t just go away and it’ll keep learning and stuff so we’re hoping to be able to really train it better on auto hotkey and have something that’s you know even better than it does to

(06:03) actually Chat GPT does amazing stuff without a hockey for for one from language it’s not really that popular it’s pretty impressive that it’s writing code as close as it gets honestly that is correct thanks everyone um talk to you soon cheers bye. Make sure you download the Chat GPT AutoHotkey tools!

🤖Unlock the Brilliant Power of ChatGPT3 API with AutoHotkey!🔥 Don’t let the future pass you by❗

🤖Unlock the Power of ChatGPT3 API with AutoHotkey!🔥

How to use ChatGPT3 API with AutoHotkey: require different parameters for web and API calls to get desired results.

00:00 🤖 Using AutoHotkey with Chat GPT and Open AI requires different parameters for web and API calls.
- We’re showing how to use AutoHotkey with Chat GPT and Open AI, and there are some important differences between the parameters used on the web version and the API calls.
01:37 🔑 Get an API token to submit input and output, based on character count.
- You need an API token to submit input and output, and the formula for calculating tokens is based on characters.
02:47 🤔 N answers are returned per request, with a limit of 4,000 tokens.
- There is a limit of 4,000 tokens per request and the number of answers it gives back is N.
03:26 🤔 Play with the API library to get the desired results, but be aware of the temperature’s effect.
- Lowering the temperature between 0 and 1 allows for factual answers, while a higher temperature of 1 allows for more creative answers.
- Play with the API library to get the desired results, but be aware of the odd behavior between temperature and something else.
05:09 🤔 ChatGPT can’t store history of its actions.
- GPT cannot send two values at the same time, so there is no history of what it does.
06:08 🤔 The rules of making a TLDR are to keep the most important events and ideas, be concise, and start with an appropriate emoji.
- Every prompt is independent and context must be manually coded into the input box.
07:19 💰 Paying for calls, but API should still work even when Chat GPT is busy and we’re separated from the free version.
- We are paying money for every call and should be able to use the API even when chat GPT is bogged down and busy, and we are likely separated from the free version.
08:35 🔥 Get 50% off our new course to learn how to use ChatGPT3 API with AutoHotkey for easier tasks!
- Check out our version 2 course on sale now for half off to learn how to use the amazing tool Chat GPT to make everything simpler.

How to use ChatGPT3 API with AutoHotkey

(3) 🤖Unlock the Power of ChatGPT3 API with AutoHotkey!🔥 – YouTube

Transcript:
(00:00) hey everyone it’s Joe and azaleas here from the automator today we got a really fun topic we’re blending Auto hotkey and also using it with chat GPT and open AI uh in this example we’re going to show you and also give you a link to get the r code for using the open ai’s API calls so this is simply you know browser call is technically an API call but this is using the web service API call and there’s some really important things to realize about it so make sure you stick around and watch the whole thing because

(00:28) there’s some gotchas at the end that we want to make sure you understand yeah that’s right and so let me show you real quick what I’ve got here so if we go ahead and type give me five words here we should get an output from chat gbt with the with this list of parameters and I think this is something that we might want to mention a little bit these parameters here are not the same as when you are in chat ai.

(01:08) com so you might get some answers from chat GPT that are not the same that you would get if you are on the web version of it so you will have to play a little bit with those guys on their website um on the documentation they really mentioned what temperature means uh what top P it is what was what is that on the end so I’m just gonna give you a very high level overview about them the maximum tokens has to do with the length of the answer and what you give it and you have to be very careful with that one because

(01:48) um as this is leveraging your token your change GPT token you might get charged if you use too many tokens right so I would rather have it in lower values if you don’t don’t want to get charged too much um but in certain situations if you want the thing to answer with 200 words you might need to bump up your maximum tokens here and let’s back up a step here just slightly so a you need to have your token you need to create an account you know and give it a credit card or something because there are cost costs

(02:24) involved unlike if there’s a browser with account the next one is once you’ve given it your token um I’m sorry you’re well it is your API token uh when you submit stuff and that’s where he’s saying there’s both the input and the output count against you yes what’s really crazy is the formula for calculating tokens it’s not like every certain characters yeah the good news is it’s actually so cheap right now at least like you you know it’s it’s you don’t really have to

(02:56) worry too much but I I still agree with what Jose is saying so you don’t want to put in a ten thousand and then suddenly realize later like hey wait a minute you know Where’d I get this bill but we’ve yet to have like any sort of a real bill that’s for real and they do have still a limit there’s 4 000 tokens per per request I think it is this is the limit that’s a kind of like a hard hard stop even if you don’t put a limit to yourself so um the N is the number of answers that it might give back right now this

(03:25) interface is just gonna display one of them so this is not really uh an option that is going to be too much flexibility because I only have one output here if I had several windows I could give you different answers right the temperature this is the one that is a little bit interesting because uh uh a 100 is kind of like like giving it freedom to be creative so it it it might come up with new stuff but sometimes you don’t want that sometimes you want it to be factual so you would lower the temperature I think

(03:59) that’s how it works and uh it is between zero and one so it’s 0.2 is a temperature of 0 of 20 and so on the same happens here with the top P so this is uh it’s just a top percent and from the answers that it generates is going to give you the top percentage and usually 0.

(04:24) 5 is okay if you want kind of like the most accurate answer it would be the top one you know the top ten you can do the top ten percent and so on so that you play with this to get the answers that you’re hoping for I wouldn’t use the stream here because that’s where it starts typing out the stuff I have it set up I but the problem the the point is I I let it here so that you know it is available as an option in the API library that we created so you take a look at it and you play a little bit with it and see what you can come up

(04:58) with you from memory there was something to me odd I don’t think they explained it very well if I remember correctly between the temperature and the what was the top p is that what you’re saying yes that yeah that yeah but not the other right exactly that’s what they said like do not change both at the same time but it is not you cannot send those two values I send them both at the same time in the code but they they they seem to only understand one of both so yeah in any case I think this is the one

(05:38) can we go back to the the GUI yes so here and let’s demonstrate it so um say please do this again please do this again now in chat GPT in the browser yeah um we would know what that would do right right but here he just answered sure because unfortunately and I have no idea why and maybe at some point we can figure out how there’s a better way to do it but there’s no history there’s no context they’re independent of each other unfortunately each input you are sending right now is independent of the previous

(06:16) one there is no context now there is a way on because if you take a look at the code I’m sending a um uh Json string with some information there is a way to set up the context in the Json string the problem is that I cannot set it up because for you to set up the context you need to get the previous um ID of the conversation and that’s the part that I haven’t really created yet but that is part of the the issue there is no context automatically you actually have to code it yourself that’s what happens so that’s something

(06:59) that you have to take into consideration when you’re using this every prompt you’re doing is an individual prompt and everything that you need as a context should be in the input box right there in the last really big gotcha which which we’re not 100 sure on when you did one test but we’re not 100 sure is theoretically to me and I know from what I’ve read too we’re paying money for every call it shouldn’t have a problem when the servers are really bogged down and busy you should be able to do this

(07:29) but one time when we went where we were using the browser and then we’re like oh it’s not running oh we’ll just use our ChatGPT3 API with AutoHotkey it also didn’t work so we’re not 100 sure if this will or won’t work when chat GPT is bogged down and busy right it probably means that we are probably separated from the free version somehow but sometimes there might be a thing with channel gbt that all of it is down so yeah that’s what I was thinking as well yeah I’m thinking that probably what we got was that we’re growing like

(08:07) crazy service was down for everybody but we are somehow separated from the free version that probably um we will not have some issues that are free person would have yeah so let us know what you think are you using chat GPT are you trying to use their API I’ll put a URL up with a lot of training on API calls which we have a lot of webinars and videos separate videos and stuff on them also this tool we’re giving you it is in version two so we did Just Launch our you know painlessly transition to V2 this is a great time to

(08:43) start learning V2 so uh in this code it’s in written in version two so you know you might want to consider our course I’ll put a URL up here it’s on sale right now for a half off so you might want to check it out but if you learned something here please like the video really helps us out we get a lot more views we’re the largest Auto hockey channel by far I noticed the other day we have over 1,400 videos and we push out two to three a week and um also don’t forget to like And subscribe because if you don’t subscribe you may

(09:12) miss out on some of our cool stuff that we’re doing later at some point we’ll you know go into this a bit more make this a little more robust to do some other stuff but ChatGPT3 API with AutoHotkey is just a crazy amazing tool that is really um making it so everything we do is so much simpler there you go that’s right thanks everyone cheers

Fine tuning Chat GPT-3 for industries Use cases

This webinar is about Fine tuning Chat GPT-3 for specific industries (and several use cases).

Chat GPT-3 is a deep learning model developed by OpenAI that can generate text for tasks such as summarization and question answering. The model can be fine-tuned to improve accuracy and performance by training on specific data sets.

This can be done by providing a reasonable dataset and using an API key from OpenAI. The fine-tuned GPT-3 model can be used for various applications such as question-answering, summarization, entity recognition and classification, and can perform tasks that previously required other machine learning models.

We share our AutoHotkey script for interacting with the Chat GPT-3 API here. You’ll still need your own account & token however this will help get you started.

Fine tuning GPT 3 for industries Use cases

Chat GPT-3 is a unidirectional deep learning model that can generate sequences for tasks such as question answering and summarizing, and can be fine-tuned to improve accuracy and performance.

00:00 🤖 We’ll discuss fine tuning Chat GPT-3 , an AI-powered neural network model developed by OpenAI that uses large amounts of data and compute power to generate sophisticated machine-generated text.
1. We are a technology-focused company harnessing AI, automation, and predictive capabilities to streamline processes and navigate the future, with three experienced speakers.
2. We will discuss training, fine-tuning, and best practices for Chat GPT-3 today.
3. Chat GPT-3 is an AI-powered neural network model developed by OpenAI that can generate large volumes of relevant and sophisticated machine-generated text with just a small amount of input.
4. Chat GPT-3 was trained on 45TB of Common Crawl data, filtered down to 570GB of plain text, and used weighted sampling to train on 410B tokens in 300B epochs.
5. Chat GPT-3 used 50 petaflop days and a ton of data to train, while RoBERTa used 50 petaflop days and less data.
6. Increasing data and compute power results in a decrease in validation loss, making for an efficient and effective training model.
15:21 🤖 Chat GPT-3 can be fine-tuned to generate higher quality results, train more examples, and save tokens with proper data and planning.
1. Chat GPT-3 can be pre-trained to perform domain-specific tasks with tuning and fine-tuning for better accuracy.
2. K-shot learning provides a way to condition a model to give the best response at inference time by giving it a few demonstrations.
3. Chat GPT-3 can generate names for superheroes based on animal behavior and complete sentences based on context from data sets like Lambda and Hella Swag.
4. Increasing the number of parameters in a model improves accuracy, but it is still not at par with the state of the art models.
5. Fine tuning Chat GPT-3 improves accuracy and widens the horizon of what problems it can solve, resulting in higher quality results, ability to train more examples, and token savings.
6. Properly training your data with the right data and planning can lead to savings and lower latency with a better response.
26:58 🤖 Fine-tuning Chat GPT-3 with a specific data set can improve its performance for multi-step reasoning problems, allowing businesses to use it for interactive question and answer sessions.
1. Fine-tuning Chat GPT-3 for a challenging multi-step reasoning problem data set resulted in significant improvement, reaching a 70 mark.
2. Keeper Tax helps users save taxes by understanding financial statement data, Viable App summarizes customer feedback to make decisions, and an LMS system generates questions and answers based on training content.
3. Fine-tuning a GPT model allows for more specific output tailored to the needs of a business, such as an interactive question and answer session for research and brainstorming.
4. GBD3 is a powerful machine learning model that can be fine-tuned to solve specific problems with the help of a dataset.
5. Publicly listed companies in the US must submit filings to the SEC, which contain a goldmine of information that cannot be answered by a generic model.
6. To fine-tune Chat GPT-3 for a question answering use case, you need to have your data set in a specific format as listed by Open AI.
36:33 🤖 Create a fine-tuned Chat GPT-3 model for question-answering by providing a reasonable dataset, using an API key from Open AI, and running a command to pass information to a server.
1. To fine-tune Chat GPT-3 for a specific document, we need to provide it with a reasonable dataset of curated paragraphs with question-answer pairs.
2. To create a fine-tuned model, you need an API key from Open AI and a file in JSONL format, which can be converted using Open AI tools or a simple code.
3. We use a command to pass information to a server, which creates a fine-tuning job and a fine-tuned model that can be used for question answering.
4. Using a fine-tuned Chat GPT-3 model, we can extract information from SEC filings for financial services, such as Facebook’s 2018 profit of 15.9 billion dollars.
5. We designed a mobile app for a foodies paradise that uses a dataset of 500 questions and answers about European countries to provide tailored information.
6. Flutter is a good framework to use for building conversational AI and video chatbot integrations due to its existing packages, well-designed integrations, and tested integrations.
46:14 🍴 Chat GPT-3 is a model used for summarization, question generation, entity classification, and entity recognition, enabling tasks that previously required other machine learning models.
1. The three popular German dishes are Grünkohl, Sauerkraut, and Rinderroulade, which are all from the Chat GPT-3 fine-tuned model.
2. Chat GPT-3 enables you to do classification and other tasks that previously required other machine learning models.
3. Chat GPT-3 is a model used to summarize domain-specific texts, such as medical abstracts, and can be used for content writing and educational purposes.
4. Chat GPT-3 can generate physics problems or questions about racing cars with the help of a few examples and fine-tuning.
5. Using Chat GPT-3, you can train a model to classify customer feedback into categories such as good, very good, bad, and very bad.
6. Chat GPT-3 helps with summarization, question generation, entity classification, and entity recognition.
51:50 🤖 Chat GPT-3 has been fine-tuned for 9 months to improve accuracy for classification and sequence generation, but it is not specifically designed for translation and PC is not ideal for regional names detection.
1. Chat GPT-3 fine tuning should be considered proportionally, as the charges depend on the number of tokens and length of the data set.
2. We have extensive experience in deploying transformer-based architectures for various NLP tasks.
3. Chat GPT-3 has been fine-tuned for the last 9 months, allowing for more accurate classification and sequence generation than GB2.
4. Chat GPT-3 is trained on English documents and can be used for translation, but it is not specifically fine-tuned for this purpose.
5. PC can be used to detect foreign names, but it does not perform well on Indian names or other regional names.
6. You need to have ground storage, databases, curation, training data, model retuning, evaluation and monitoring to create an end-to-end workflow for an organization’s requirements.
59:20 🤔 Chat GPT-3 is a unidirectional deep learning model that can generate sequences for tasks such as question answering and summarizing, but requires human help for more complex tasks like sentiment analysis.
1. It is possible to match intent and meaning of two different questions using simpler models such as cosine similarity, but for more complex tasks such as sentiment analysis, a human being must help a deep learning model like Chat GPT-3.
2. The metric used to identify if more prompt is needed to improve accuracy varies depending on the problem and type of classification.
3. Chat GPT-3 can be used for classification problems and can be used with fixed sets of values to generate intent without needing to be fine-tuned.
4. Chat GPT-3 is unidirectional, meaning it is dependent on context on one side, which is both a strength and limitation.
5. Chat GPT-3 is explicitly trained to focus only on the left context and mask the right context, making it unidirectional.
6. Chat GPT-3 is a generative model that can understand context and generate sequences for tasks such as question answering and summarizing.
01:09:34 🤖 Chat GPT-3 has expanded the range of machine learning problems that can be solved, and it is easy to integrate and fine-tune with only a few samples.
1. When using the same prompt, results are generally different, but seeding and providing examples can help control the output.
2. You can control the randomness of Chat GPT-3 using temperature, and use cosine similarity to measure semantic similarity without Chat GPT-3.
3. Language models are often fine-tuned on Eurocentric datasets, but there is an opportunity to use GBC to direct data and model drift for other parts of the world.
4. Chat GPT-3 has expanded the range of machine learning problems that can be solved, and it is easy to integrate and fine-tune with only a few samples.
5. For generative use cases, fine tuning of lower models is recommended as it costs less and requires fewer tokens than higher models.
6. Pass parameters to control model name, GPU type, and other things when using GPT3 and MLPs.