At its first DevDay, OpenAI launched GPT-4 Turbo with a 128,000-token context window and lower prices, alongside no-code customizable GPTs, an upcoming GPT Store, and the Assistants API for building agents.
OpenAI starts rolling out voice and image capabilities in ChatGPT. The assistant can now hold spoken conversations in five different voices and analyze photos thanks to GPT-4V.
OpenAI unveils DALL-E 3, its new image-generation model, built directly into ChatGPT. The key novelty: it's controlled with natural language and lets you iterate on an image within the conversation itself.
George R. R. Martin, John Grisham and 17 other authors have joined the Authors Guild in a class-action lawsuit accusing OpenAI of using their books without permission to train ChatGPT.
OpenAI unveils a corporate version of ChatGPT with stronger security, unlimited GPT-4 access and a context window four times larger. The company says it won't use customer data to train its models.
OpenAI has introduced function calling to its API, letting GPT-4 and GPT-3.5-turbo generate JSON arguments to invoke external functions—a key step toward agents that act on other systems.
The era of autoregressive neural networks has marked a turning point in natural language processing (NLP). Among the most significant developments in this ar...
Artificial Intelligence (AI) has seen an exponential growth in its capabilities, thanks to the advancement of language models such as GPT-3, BERT, and more r...
Recent advances in artificial intelligence (AI) have propelled large-scale language models, such as GPT-3 and BERT, to the forefront of natural language proc...
GPT (Generative Pre-trained Transformer) is not a model but a family of OpenAI systems built on a decoder-only transformer and pre-trained in a self-supervised way. A tour of its method and lineage, from GPT-1 to today's family, without crowning any single version.
Elon Musk has incorporated X.AI Corp, a new company targeting the artificial intelligence market. The move comes after ChatGPT’s success and as Musk calls for a pause in the development of systems more powerful than GPT-4.
T5, from Google (Raffel et al., 2019), treats every language task as a “text-to-text” problem. We explain that unified approach, its encoder-decoder architecture versus GPT and BERT, its pre-training on the C4 corpus, and variants such as mT5 and Flan-T5.
This website uses cookies to improve the browsing experience. Cookie policy.
No results. Try different words.
