𝗧𝗵𝗶𝘀 𝗶𝘀 𝗵𝗮𝗻𝗱𝘀 𝗱𝗼𝘄𝗻 𝗼𝗻𝗲 𝗼𝗳 𝘁𝗵𝗲 𝗕𝗘𝗦𝗧 𝘃𝗶𝘀𝘂𝗮𝗹𝗶𝘇𝗮𝘁𝗶𝗼𝗻 𝗼𝗳 𝗵𝗼𝘄 𝗟𝗟𝗠𝘀 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝘄𝗼𝗿𝗸. ⬇️ 𝘓𝘦𝘵'𝘴 𝘣𝘳𝘦𝘢𝘬 𝘪𝘵 𝘥𝘰𝘸𝘯: 𝗧𝗼𝗸𝗲𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻 & 𝗘𝗺𝗯𝗲𝗱𝗱𝗶𝗻𝗴𝘀: - Input text is broken into tokens (smaller chunks). - Each token is mapped to a vector in high-dimensional space, where words with similar meanings cluster together. 𝗧𝗵𝗲 𝗔𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻 𝗠𝗲𝗰𝗵𝗮𝗻𝗶𝘀𝗺 (𝗦𝗲𝗹𝗳-𝗔𝘁𝘁𝗲𝗻𝘁𝗶𝗼𝗻): - Words influence each other based on context — ensuring "bank" in riverbank isn’t confused with financial bank. - The Attention Block weighs relationships between words, refining their representations dynamically. 𝗙𝗲𝗲𝗱-𝗙𝗼𝗿𝘄𝗮𝗿𝗱 𝗟𝗮𝘆𝗲𝗿𝘀 (𝗗𝗲𝗲𝗽 𝗡𝗲𝘂𝗿𝗮𝗹 𝗡𝗲𝘁𝘄𝗼𝗿𝗸 𝗣𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴) - After attention, tokens pass through multiple feed-forward layers that refine meaning. - Each layer learns deeper semantic relationships, improving predictions. 𝗜𝘁𝗲𝗿𝗮𝘁𝗶𝗼𝗻 & 𝗗𝗲𝗲𝗽 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 - This process repeats through dozens or even hundreds of layers, adjusting token meanings iteratively. - This is where the "deep" in deep learning comes in — layers upon layers of matrix multiplications and optimizations. 𝗣𝗿𝗲𝗱𝗶𝗰𝘁𝗶𝗼𝗻 & 𝗦𝗮𝗺𝗽𝗹𝗶𝗻𝗴 - The final vector representation is used to predict the next word as a probability distribution. - The model samples from this distribution, generating text word by word. 𝗧𝗵𝗲𝘀𝗲 𝗺𝗲𝗰𝗵𝗮𝗻𝗶𝗰𝘀 𝗮𝗿𝗲 𝗮𝘁 𝘁𝗵𝗲 𝗰𝗼𝗿𝗲 𝗼𝗳 𝗮𝗹𝗹 𝗟𝗟𝗠𝘀 (𝗲.𝗴. 𝗖𝗵𝗮𝘁𝗚𝗣𝗧). 𝗜𝘁 𝗶𝘀 𝗰𝗿𝘂𝗰𝗶𝗮𝗹 𝘁𝗼 𝗵𝗮𝘃𝗲 𝗮 𝘀𝗼𝗹𝗶𝗱 𝘂𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴 𝗵𝗼𝘄 𝘁𝗵𝗲𝘀𝗲 𝗺𝗲𝗰𝗵𝗮𝗻𝗶𝗰𝘀 𝘄𝗼𝗿𝗸 𝗶𝗳 𝘆𝗼𝘂 𝘄𝗮𝗻𝘁 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱 𝘀𝗰𝗮𝗹𝗮𝗯𝗹𝗲, 𝗿𝗲𝘀𝗽𝗼𝗻𝘀𝗶𝗯𝗹𝗲 𝗔𝗜 𝘀𝗼𝗹𝘂𝘁𝗶𝗼𝗻𝘀. Here is the full video from 3Blue1Brown with exaplantion. I highly recommend to read, watch and bookmark this for a further deep dive: https://lnkd.in/dAviqK_6 𝗜 𝗲𝘅𝗽𝗹𝗼𝗿𝗲 𝘁𝗵𝗲𝘀𝗲 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁𝘀 — 𝗮𝗻𝗱 𝘄𝗵𝗮𝘁 𝘁𝗵𝗲𝘆 𝗺𝗲𝗮𝗻 𝗳𝗼𝗿 𝗿𝗲𝗮𝗹-𝘄𝗼𝗿𝗹𝗱 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲𝘀 — 𝗶𝗻 𝗺𝘆 𝘄𝗲𝗲𝗸𝗹𝘆 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿. 𝗬𝗼𝘂 𝗰𝗮𝗻 𝘀𝘂𝗯𝘀𝗰𝗿𝗶𝗯𝗲 𝗵𝗲𝗿𝗲 𝗳𝗼𝗿 𝗳𝗿𝗲𝗲: https://lnkd.in/dbf74Y9E | 48 comments on LinkedIn

OpenAI Modelle auf der Website ganz einfach „finetunen“ und damit eigene Versionen des ChatGPT-Modells erstellen.

🚀 🚀 🚀 #Orca is a language model that is designed to solve the limitations of smaller models by imitating the reasoning processes of #GPT-4. As Orca is…

Die rasante Entwicklung generativer KI-Systeme wie ChatGPT, DALLE und Midjourney revolutioniert unsere Welt und stellt uns vor neue, faszinierende und zugleich beunruhigende Herausforderungen. Was bedeutet es, wenn künstliche Intelligenzen komplexe Prüfungen bestehen oder sogar kreativ tätig werden? Wird diese Entwicklung unsere Gesellschaft, Arbeitswelt und Kommunikation überrumpeln? Werden Jobs ersetzt oder entstehen völlig neue Beschäftigungsfelder? Und wie können wir die Gefahren durch die Nutzung dieser Systeme zur Verbreitung von z.B. Falschinformationen minimieren? Entdecken Sie in diesem vierwöchigen, kostenlosen openHPI-Kurs, wie bahnbrechende Technologien wie ChatGPT funktionieren, welche Anwendungsfälle daraus entstehen, welche Chancen und Grenzen sie bergen. Der Kurs bietet Jugendlichen und Interessierten ohne technisches Hintergrundwissen oder Programmiererfahrung eine einzigartige Gelegenheit, in die Welt der Generativen Künstlichen Intelligenz einzutauchen. Geleitet wird der Kurs in Kooperation mit dem [KI-Servicezentrum Berlin-Brandenburg (KISZ-BB)][2] von den HPI-Alumni [Johannes Hötter][3] und [Christian Warmuth][4]. [1]: https://open.hpi.de/courses/kieinstieg2020 [2]: https://hpi.de/forschung/infrastruktur/ki-servicezentrum-berlin-brandenburg.html [3]: https://www.linkedin.com/in/johannesh%C3%B6tter/ [4]: https://www.linkedin.com/in/warmuth-christian/

Greg Brockman von OpenAI skizziert neue Nutzungsszenarien für ChatGPT und die Zusammenarbeit mit smarten Maschinen.

Ask any AI chatbot a question, and its answers are either amusing, helpful, or just plain made up. Here's how to prevent AI hallucinations.

Gen AI is evolving at record speed while CEOs are still learning the technology’s business value and risks. Here, we offer some of the generative AI essentials.

Discover the key differences between AutoGPT and HuggingGPT, two leading AI language models based on GPT-4, as we compare their strengths, weaknesses, and pe...

Open Source will win over other LLM. You can build on Bing your apps!!! Breaking down all the AI News from the week. This week really started picking up speed again. 🛠️ Explore hundreds of AI Tools: https://futuretools.io/ 📰 Weekly Newsletter: https://www.futuretools.io/newsletter 😊 Discord Community: https://futuretools.io/discord 🐤 Follow me on Twitter: https://twitter.com/mreflow 🐺 My personal blog: https://mattwolfe.com/ 🌯 Buy me a burrito: https://ko-fi.com/mattwolfe --- Links From The Video: https://www.theverge.com/2023/5/1/23706311/hinton-godfather-of-ai-threats-fears-warnings https://finance.yahoo.com/news/ibm-pause-hiring-plans-replace-212747073.html https://www.cnbc.com/2023/05/02/chegg-drops-more-than-40percent-after-saying-chatgpt-is-killing-its-business.html https://www.vice.com/en/article/4a3w3g/scientists-use-gpt-ai-to-passively-read-peoples-thoughts-in-breakthrough https://www.reuters.com/technology/reid-hoffmans-new-ai-startup-inflection-launches-chatgpt-like-chatbot-2023-05-02/ https://inflection.ai/ https://heypi.com/talk https://twitter.com/Mr_AllenT/status/1653753737888821248 https://nftnow.com/features/how-a-digital-artist-used-chat-gpt-to-create-a-50m-meme-coin/ https://www.youtube.com/watch?v=RI_dWnaUuw4 https://www.theverge.com/2023/5/4/23710533/google-microsoft-openai-white-house-ethical-ai-artificial-intelligence https://www.semianalysis.com/p/google-we-have-no-moat-and-neither https://blogs.microsoft.com/blog/2023/05/04/announcing-the-next-wave-of-ai-innovation-with-microsoft-bing-and-edge/ --- Outro music generated by Mubert https://mubert.com/render #AINews #AITools #openai --- 0:00 Intro 0:21 The Godfather of AI 0:47 IBM Replacing 7,800 Jobs 1:42 Chegg is Getting Destroyed 2:05 AI Can Read Your Mind 3:46 Inflection AI Creates PI 5:08 MidJourney V 5.1 6:43 AI Creates $50m Crypto Coin 8:35 AI Leaders Meet At White House 9:48 Google is Getting Scared 12:27 Microsoft's Bing Chat For ALL 14:57 Future Tools

Habe auf dem Data Science Summit 2023 einen Vortrag über die Auswirkungen von ChatGPT / Generative AI auf Wirtschaft und Arbeit gehalten. Bisher sehen wir erst…

Das Paper gibt eine Überblick über ChatGPT/NLP-Anwendungen, vor allem in wissensintensiven Cases. Nach einer Studie arbeiten inzwischen 53 Prozent der Data…