OpenAI Launches Enhanced GPT-4 Turbo and New Embedding Models: Addressing ‘Laziness’ and Expanding Capabilities

OpenAI Launches Enhanced GPT-4 Turbo and New Embedding Models: Addressing ‘Laziness’ and Expanding Capabilities. OpenAI has enhanced GPT-4 with new updates, addressing the issue of ‘laziness’ in task completion. The improved version, known as GPT-4 Turbo, is said to perform tasks, particularly code generation, more effectively compared to its predecessor. However, the specifics of these updates were not disclosed by the company.

Recently, some users of ChatGPT reported frequent refusals by the chatbot to complete tasks, attributing this to the absence of updates in GPT-4. OpenAI’s recent update targets GPT-4 Turbo, an advanced variant of the widely used GPT-4, which incorporates data up to April 2023 and is presently in preview mode. Users of the standard GPT-4, trained on data up to September 2021, might still encounter similar performance issues.

According to OpenAI’s blog post, over 70% of GPT-4 API users have switched to GPT-4 Turbo, drawn by its more contemporary knowledge base. The company also announced upcoming enhancements for GPT-4 Turbo in the following months, including the addition of vision capabilities to the general release. This will enable users to execute more complex multimodal prompts, such as text-to-image generation.

OpenAI Launches Enhanced GPT-4 Turbo and New Embedding Models: Addressing ‘Laziness’ and Expanding Capabilities

In addition, OpenAI has introduced new, smaller AI models called embeddings. Defined as a series of numbers representing the concepts in content like natural language or code, embeddings aid applications that use retrieval-augmented generation—a form of AI that relies on database information rather than generating answers independently—to understand the connections between different content types.

Two versions of these models, the less powerful text-embedding-3-small and the more robust text-embedding-3-large, are now available. Furthermore, OpenAI has announced two additional updates: an enhanced preview of GPT-4 Turbo with more efficient task completion and a resolution for a UTF-8 encoding issue, along with the release of a new GPT-3.5 Turbo next week, featuring a 50% reduction in input prices and a 25% reduction in output prices.

Related Posts:

What is Chat GPT AI/ML?(Opens in a new browser tab)

Automotive company Magna invests $5 million in AI systems(Opens in a new browser tab)

Toronto Consolidates IT Systems for Enhanced Security Following Ransomware Attacks on Zoo and Public Library(Opens in a new browser tab)

How to sign into and use ChatGPT(Opens in a new browser tab)

Elon Musk and scientists ask for pause to AI ChatGPT(Opens in a new browser tab)

Electric Vehicle Challenges: Navigating the Complex Landscape of EV Marketing and Consumer Skepticism(Opens in a new browser tab)