- OpenAI’s GPT-3 models, powering Bing Chat and ChatGPT, are performing well in tests designed to assess cognition in children.
- ChatGPT is showing performance similar to that of a 9-year-old.
ChatGPT performs like a 9-year-old in ‘theory of mind’ test

- OpenAI’s GPT-3 models, powering Bing Chat and ChatGPT, are performing well in tests designed to assess cognition in children.
- ChatGPT is showing performance similar to that of a 9-year-old.
[Via]