Anthropic, a startup founded by former members of OpenAI, has released the beta version of Claude 2, the latest version of its chatbot aimed at rivaling OpenAI’s ChatGPT. Claude 2 offers longer responses, improved performance, and better safety guardrails compared to its predecessor. The chatbot performed well on a bar exam and the GRE, and it allows users to input hundreds of pages for more efficient summarization. Claude 2 is available for beta testing through a public-facing website and an API in the US and the UK. While it demonstrates proficiency in addressing common-sense questions, simple math, and document summarization, its writing quality could still improve when it comes to qualitative queries, exhibiting repetitive word choices and rambling answers that plague many other broadly trained models.
Why it matters: With its improved performance, longer responses, and enhanced safety guardrails, Claude 2 introduces a viable alternative for users seeking advanced conversational capabilities and efficient document summarization.
- Anthropic has raised over $1 billion in funding to develop a large language model (LLM) to rival ChatGPT.
- Claude 2 offers improved performance, longer responses, and better safety guardrails compared to its previous iteration. It also allows users to input hundreds of pages of text, enabling more efficient summarization of complex documents.
- The beta introduces a viable alternative for users seeking advanced conversational capabilities and efficient document summarization.
