Although large language models (LLMs) are incredibly capable, they are pretty simple to understand. In fact, the core components of most LLMs can be distilled into five major components… 🧵[1/7]
here is GPT-4, our most capable and aligned model yet. it is available today in our API (with a waitlist) and in ChatGPT .
openai.com/research/gpt-4
it is still flawed, still limited, and it still seems more impressive on first use than it does after you spend more time with it.
Das Didaktik-Medien-Zentrum (DMZ) der #HS_Augburg hat jetzt eine eigene Seite auf LinkedIn - gern folgen, wer sich für Didaktik, Hochschullehre, digitale/hybride Lehre interessiert linkedin.com/feed/update/urn…
Just made ChatGPT available on our API! Our incredible team of builders has delivered 10x cheaper model than our existing GPT-3.5 models, through system-wide optimizations, making it easier to power as many applications as possible.
openai.com/blog/introducing-…
Today is a huge day for developers. 🤯
- ChatGPT API released (10x cheaper)
- Whisper available in the API
- Overhauled data usage policy
- Focus on stability
And more!
Here’s a quick thread on everything we shipped today @OpenAI 🧵
Warum erzählt die @fdp immer und immer wieder das Märchen der #eFuels und droht auch noch damit das #EU-#Verbrenner-#Verbot zu blockieren? Eine Verkehrspolitik von vorgestern bedroht am Ende auch die Automobilindustrie in 🇩🇪.
Some personal news: I am joining OpenAI (again :)). Like many others both in/out of AI, I am very inspired by the impact of their work and I have personally benefited greatly from it. The future potential is especially exciting; it is a great pleasure to jump back in and build!🪄
1/ In 2021, we shared next-gen language conversation capabilities powered by our Language Model for Dialogue Applications (LaMDA). Coming soon: Bard, a new experimental conversational #GoogleAI service powered by LaMDA.
blog.google/technology/ai/ba…
I tried to make a super compact video about how to use ChatGPT to train a neural network in a web-based environment (Google Colab) with relatively little prior knowledge. #deeplearning#chatgpt#python#jupyteryoutu.be/xpFK6ncez6s
Incredible paper. LLMs-generated text can be detected by embedding signals that are invisible to humans but algorithmically detectable.
It's a watermarking framework for your language models.
Reaches 99% confidence on 23 words.
📄arxiv.org/abs/2301.10226
🧠@tomgoldsteincs
🔥 New (1h56m) video lecture: "Let's build GPT: from scratch, in code, spelled out."
youtube.com/watch?v=kCc8FmEb…
We build and train a Transformer following the "Attention Is All You Need" paper in the language modeling setting and end up with the core of nanoGPT.
Everyone is amazed at OpenAI and ChatGPT, but don’t forget: the transformer was invented by Google, the first LLM was Google’s BERT, and Google made an apparently impressive chatbot (LaMDA) before ChatGPT but didn’t release publicly since they didn’t feel it was safe to do so.