Claude AI benefits from significant improvements with its latest update, what should ChatGPT be concerned about?
While OpenAI plays to sabotage itself, the competition is active. Anthropic, one of the major players in generational artificial intelligence, has just announced Claude 2.1, the latest version of its chatbot and major competitor to the famous ChatGPT.
Twice the processing capacity
One of the big new features of this update is the doubled limit on the amount of information that can be transmitted to Claude, which now supports up to 200,000 tokens (for Claude Pro only, the version paid). This amounts to approximately 150,000 words, or over 500 pages of text.
Users can thus submit extremely long documents, such as substantial literary works, such as The Iliad Or The Odysseytechnical notices, code bases, legal texts… Anthropic nevertheless warns that processing such a quantity of data can take time, several minutes, but that performance should improve over time.
As before, on smaller quantities of data, Claude can summarize in the form of your choice, generate questions and answers, predict trends, or even compare several documents.
More reliable and more complete
With Claude 2.1, hallucination rates are halved, Anthropic claims. This means that it will generate false information half as often as Claude 2.0. All language models today suffer from this phenomenon, with responses that are sometimes completely invented and we do not really understand what they are based on.
The company also mentions “significant improvements in comprehension and synthesis, particularly for long and complex documents that require a high degree of precision, such as legal documents, financial reports and technical specifications”. In this context, Claude 2.1 benefits from a 30% reduction in incorrect answers and an incorrect conclusion rate 3 to 4 times lower than the previous version.
Finally, Anthropic deployed in beta a chatbot interoperability function with third-party APIs, allowing it to more easily access certain tools and data. Claude can therefore use a calculator, access databases, perform a web search, perform simple actions in software via private APIs or connect to product datasets to make recommendations.
Claude 2.1 is available both within the API and in the user interface, for paying subscribers or Internet users who use the free version.
Download
- Upload files of up to 100,000 tokens (approximately 75,000 words)
- Advanced customization
- Ethical design
Claude 2, the first born of the Anthropic family, is a chatbot that attempts to push the limits of artificial intelligence. It is used to facilitate digital interactions, automate tasks and provide precise and contextual responses. With the capacity to process up to 100,000 tokens in a single request, Claude 2 is a powerful tool for those looking to optimize their digital workflow.
Claude 2, the first born of the Anthropic family, is a chatbot that attempts to push the limits of artificial intelligence. It is used to facilitate digital interactions, automate tasks and provide precise and contextual responses. With the capacity to process up to 100,000 tokens in a single request, Claude 2 is a powerful tool for those looking to optimize their digital workflow.
Source : Anthropic
10