GlobalTechnoNews
- AI in Business and Industry , AI Research and Development , Artificial Intelligence (AI)
- October 16, 2024
- 250 views
Apple: Artificial Intelligence Isn’t as Smart as It Seems
Large language models (LLMs) have shown significantly lower quality results when researchers used tests requiring logical reasoning with understanding. Large language models (LLMs) from OpenAI, Google, Meta Platforms, and others might not be as intelligent as they appear. Apple researchers discovered that these LLMs rely more on ‘sophisticated pattern matching’ than ‘true logical reasoning.’ The most common benchmark for reasoning skills is the GSM8K test. However, due to its popularity, there’s a risk of data contamination. This means LLMs…
Read more
