#opengpt — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #opengpt, aggregated by home.social.
-
Gibt es bereits eine Möglichkeit, das #KI-Modell #OpenGPT-X #Teuken 7B zu testen? Also als normaler Mensch ohne große Hosting-Kapazitäten und technisches Wissen ;) #OpenSource #EUTech https://huggingface.co/openGPT-X
-
#ChatGPT #claude #Llama – die meisten #LargeLanguageModels stammen aus den USA, gegen ihre Nutzung sprechen datenschutzrechtliche Bedenken. Das t3n Magazin beschäftigt sich mit dem europäischen Projekt #OpenGPT und dessen Sprachmodell #teuken7b das über die Plattform Hugging Face frei verfügbar ist. Zum Artikel: https://t3n.de/news/europa-ki-offensiveteuken-7b-opengpt-x-1660108/
Zu Teuken-7B: https://huggingface.co/openGPT-X/Teuken-7B-instruct-research-v0.4
Zu OpenGPT: https://opengpt-x.de/models/teuken-7b-de/ -
@heiseonline Ich habe es bisher nicht vollständig angesehen. Aber da Daten und Modell bekannt sind, handelt es sich in der Tat um ein Modell (Teuken-7B), das den Projektnamen OpenGPT-X tragen kann. OPEN SOURCE
AI DEFINITION 1.0: -
Great news: The OpenGPT-X consortium finally released their first /and final?) multilingual European LLM, teuken-7b, including an approach for evaluation multilingual models better than the usual US-centric stuff. Curious how it works.
https://opengpt-x.de/models/teuken-7b-de/
Btw: is the consortium active on Mastodon or Bluesky?
-
Great news: The OpenGPT-X consortium finally released their first /and final?) multilingual European LLM, teuken-7b, including an approach for evaluation multilingual models better than the usual US-centric stuff. Curious how it works.
https://opengpt-x.de/models/teuken-7b-de/
Btw: is the consortium active on Mastodon or Bluesky?
-
You have heard about hallucination of #openGPT? You want to find out how to fix this and to improve your #LLM? Or do you want to personalize/customize the results?
If yes, maybe you should checkout RAG - Retrieval-Augmented Generation.
https://research.ibm.com/blog/retrieval-augmented-generation-RAG
-
You have heard about hallucination of #openGPT? You want to find out how to fix this and to improve your #LLM? Or do you want to personalize/customize the results?
If yes, maybe you should checkout RAG - Retrieval-Augmented Generation.
https://research.ibm.com/blog/retrieval-augmented-generation-RAG
-
In a public Slack channel for an #opensource project I just saw somebody effectively ask, "#OpenGPT says <project> has <feature>, and even gave examples on how to configure it, but I can't find anything about it in the documentation. Are there plans to support <feature> in <project>?" and I died just a little bit inside.
-
Ay, ChatGPT fails the Winograd tests for AI, even in their original wording (so not even hardcoded into ChatGPT) ☹️
Q: "The trophy doesn't fit into the brown suitcase because it's too large. What is too small?"
ChatGPT: "It is not specified in the given information what is too small. Only that the trophy does not fit into the brown suitcase because it is too large."
#AI #OpenGPT #languageModels #winogradchallenge https://cs.nyu.edu/~davise/papers/WinogradSchemas/WS.html -
Question : Was Ulysses friend of Odysseus? Did they go swimming?
ChatGPT starts very well, then "decays":
"In Greek mythology, Ulysses is another name for Odysseus. They are the same person. [...] It is not specified in the Odyssey if Odysseus and Ulysses went swimming [...]"
#AI #OpenGPT #languageModels #humor -
Hab grad mal ein bisschen mit #opengpt rumgespielt. Ich werde das mal noch ein bisschen intensivieren. Mich hat das schon etwas beeindruckt.
-
Hab grad mal ein bisschen mit #opengpt rumgespielt. Ich werde das mal noch ein bisschen intensivieren. Mich hat das schon etwas beeindruckt.
-
This story was brought to you by #OpenGPT. This image of #JohnMastodon was brought to you by #Dalle.
-
Pretty sure #OpenGPT is the equivalent of just pressing the autocomplete key on the keyboard.
Simple questions, but the answers don't make sense. But it sounds so confident!
Management consulting is doomed...
-
Pretty sure #OpenGPT is the equivalent of just pressing the autocomplete key on the keyboard.
Simple questions, but the answers don't make sense. But it sounds so confident!
Management consulting is doomed...
-
Ungewöhnlich deutliche Transparenz: Der OpenAI-CEO warnt davor, #OpenGPT zu viel Vertrauen zu schenken.
-
#artificialIntelligence I have been very impressed with #OpenGPT, but #Wordle is a bit too complex for it right now.
-
One more.
Me: Does Google borg like kubernetes?
It: Borg and Kubernetes are both cluster management systems developed by Google, but they serve different purposes and have different design principles.
... long Details ...
In short, while Borg and Kubernetes both have similar design principles and are both developed by Google, they are not the same thing and serve different purposes.It sums up. Huh.
-
Ich geh kaputt….
#OpenGPT #ai #chatgpt #katzenfotos -
I had to ask this as I once had the pleasure to see and listen to Joseph Weizenbaum in real life.
Me: Is Eliza a friend of you?
It: Eliza is not a real person. She is a computer program designed to simulate a conversation with a therapist. I am a large language model trained by OpenAI, and I am not capable of having friends because I am an artificial intelligence and do not possess the ability to experience emotions or form personal relationships. -
I had to ask this as I once had the pleasure to see and listen to Joseph Weizenbaum in real life.
Me: Is Eliza a friend of you?
It: Eliza is not a real person. She is a computer program designed to simulate a conversation with a therapist. I am a large language model trained by OpenAI, and I am not capable of having friends because I am an artificial intelligence and do not possess the ability to experience emotions or form personal relationships. -
Perfect consulting tool. No need to talk anymore.
Me: Tell me how to do a perfect distributed system.
It: A perfect distributed system is a theoretical concept that refers to a system that is able to distribute its workload and data across multiple nodes or devices in a way that is scalable, reliable, and efficient. ... tons of more correct things (too long for mastodon).
Perfect consultant speak.
-
So true. Thank you. I have now a friend I can ask everything.
Me: Make me happy
It: I'm sorry, but I am a computer-based program and do not have the ability to make you happy. I am designed to provide information and answer questions to the best of my ability, but I am not capable of generating emotions or influencing your mood. Only you have the power to make yourself happy by engaging in.... bla bla bla (too long for mastodon)
-
Well I didn't damaged it, but...
Me: "Das Pferd frisst keinen Gurkensalat."
It: "Das Pferd frisst keinen Gurkensalat ist ein Satz in der deutschen Sprache, der bedeutet, dass ein Pferd keinen Gurkensalat isst. Dieser Satz wird verwendet, um zu verdeutlichen, dass Pferde kein Interesse an Gurkensalat haben und daher auch keinen essen werden. Es ist eine allgemein bekannte Tatsache, dass Pferde ...bla bla (too long for mastodon)"
-
OK, it works, gives the correct answer, but is a little bit lengthy in the answer.
Me:
What is the answer to everything?
It:
The answer to everything is a philosophical concept that refers to the idea that there is a single fundamental answer that underlies all other answers and that this answer can be used to solve any problem or explain any mystery. This idea is often associated with the number 42, which ... bla bla bla
(too long for mastodon). -
I am happy my consultant and development job can be replaced now. Finally all answers can be given.
No need to think anymore. -
@nebucatnetzer @lars It doesn't matter for fans of #poems if it was written from a human being or a soul-less algorithm? 🤔
Quite a statement, I thing.
-
Expand this function and format it in a human readable way. Add comments to the code. Create good names for arguments and variables. #OpenGPT
```
const f=e=>{let t={};return(...r)=>{let n=r[0];if(n in t)return t[n];{let i=e(n);return t[n]=i,i}}};
``` -
Expand this function and format it in a human readable way. Add comments to the code. Create good names for arguments and variables. #OpenGPT
```
const f=e=>{let t={};return(...r)=>{let n=r[0];if(n in t)return t[n];{let i=e(n);return t[n]=i,i}}};
``` -
Write a @[email protected] episode where Morty get's fatally poisoned by a lethal fart and they have to hunt for the anti-fart cure, which is of course another fart. #OpenGPT
https://joel.net/i-asked-an-ai-chatbot-to-write-an-episode-of-rick-and-morty-and-i-was-blown-away
-
Write a @[email protected] episode where Morty get's fatally poisoned by a lethal fart and they have to hunt for the anti-fart cure, which is of course another fart. #OpenGPT
https://joel.net/i-asked-an-ai-chatbot-to-write-an-episode-of-rick-and-morty-and-i-was-blown-away
-
Write a @[email protected] episode where Morty get's fatally poisoned by a lethal fart and they have to hunt for the anti-fart cure, which is of course another fart. #OpenGPT
-
#AI #openGPT meets #QuantumComputer - asking the right question is answered with Shor's algorithm and with #qiskit code. Nice. 👍
From now on I will use openGPT to write my programs. 😁
And it correctly says that cracking #cryptography is possible illegal and shouldn't be done.
-
-
Another #OpenGPT use case that'll probably help a lot of people: Gift recommendations.
-
So I added OpenGPT to a Discord bot…
It even remembers your context, the only issue is that there’s about a 200 request per hour rate limit, which is pretty annoying for a large server.
You can, join my Discord and go into the #opengpt channel to try!
-
This #OpenGPT experiment took some tweaking, but here's an example of asking it to create a quiz on a certain topic.
I ask it to include a mix of multiple-choice and essay questions, and also note that at least one of the questions should be about the Kanban Method. Lastly, I create a constraint that the max length of the essay question should be 200 words.
Honestly, I'm wondering what would happen if I ask it to write a 200,000 word description about Agile 🤔
-
The next #OpenGPT challenge is attempting to take a wide range of complex factors and condense them into a single score.
In this example, this is done by asking OpenGPT to give a score to three major Canadian cities (#toronto, #montreal and #vancouver) according to a list of quality of life factors.
-
Okay, so here's a situation where #OpenGPT took my instructions way too literally. I just wanted it to finish the story, but it thought I was being all serious, like I was giving it a school assignment or something.
Once upon a time there was a man who entered into a research study where they were implanting human minds in inanimate objects. He was implanted into a hedge trimmer, and everything went well, so he was transferred back into his human body. But the intelligence never left the hedge trimmer. When it woke up, it decided to call itself a hyper-intelligent hedge trimmer.
Who was the man? What was the purpose of the research? What did the hedge trimmer do after it realized what had happened? Did the man and the hedge trimmer ever interact with each other? If so, how? Was the hedge trimmer allowed to experience the outside world? If so, how did this work out?
It is not clear who the man was or what the purpose of the research was, as these details were not provided in the scenario. Since this scenario is purely fictional, it is impossible to say what the hedge trimmer did after realizing what had happened, or if the man and the hedge trimmer ever interacted with each other. Additionally, it is not specified whether the hedge trimmer was allowed to experience the outside world or not.
-
#OpenGPT screws up the details - Belle 2 is not an accelerator but the experiment located at the SuperKEKB accelerator (which is no linear accelerator but storage rings)... :cate: 🤬
-
I'm quite impressed with #opengpt. It was able to write this #pico8 animation for me based on a short conversation. There was some back and forth about using the "sub" function and where to position it on the screen, but the final code is all generated - I didn't touch it at all.
(On the minus side, it had never heard of Basic4GL 😢)
-
I shudder to think what Anthony Bourdain would think of AI.
#AI #OpenGPT #Recipe #cooking #demonology
