Claude Chatbot by Anthropic AI system- similar to ChatGPT like language model | AI Models

Have you heard of another Language model that is better than ChatGPT in some ways?

Yes, we have another Language model developed in Anthropic by ex-employees of Open AI called Claude.

They have raised $700 million to date. It is only accessible through a slack integration as it is still in the testing phase.

People who have access to it have said that it is better than ChatGPT in some ways.

Anthropic Claude is being developed by a technique called “Constitutional AI.”
The company recently said that “constitutional AI” aims to provide a “principle-based” approach to align AI with human intentions, letting AI respond to questions using a set of principles as a guide.

They used ten guiding principles to engineer Claude, collectively forming a constitution called “Constitutional AI.”

Though the ten principles are not yet public, the company has said that it is grounded on non-maleficence(avoiding giving harmful advice), beneficence(maximizing positive impact), and autonomy(freedom of speech)

Claude is developed by employing another AI system, not Claude itself, to train the system on the principles by giving inputting different prompts, exploring responses to those prompts, and curating those that are apt to the constitution, which is then packaged into a single model.

This method was then used to train Claude.

Many prompts have been tested to determine the difference between Claude and ChatGPT. Though it is found that Claude is better in some fundamental ways, it is not entirely better than ChatGPT.


#ai #chatgpt #chatgpt3 #chatgptexplained #anthropic #openai #artificialintelligence #artificialintelligencecoachingcenter