Saturday, February 15, 2025
HomeTechnologyAnthropic, OpenAI, And Meta

Anthropic, OpenAI, And Meta

If the top artificial intelligence (AI) models in the tech sector had superlatives, Microsoft-backed OpenAI’s GPT-4 would be best at maths, Meta’s Llama 2 would be most pragmatic, Anthropic’s Claude 2 would be best at understanding its limitations, and Cohere AI would be crowned as having the most hallucinations—and confidently incorrect answers.

All of this is stated in a report published on Thursday by experts at Arthur AI, a platform for monitoring machine learning.
The study was conducted at a time when generative AI is experiencing a surge ahead of the 2024 U.S. presidential election and artificial intelligence-related misinformation is being discussed more passionately than ever.

According to Adam Wenchel, co-founder and CEO of Arthur, “rather than just sort of… offer a single number that talks about where they are on an LLM leaderboard,” this is the first report to “take a full look at rates of hallucination.”

Source (CNBC)

- Advertisment -

Most Popular

Recent Comments