Table 3 Correlations of chatbots and expert ratings.

From: Large language models can outperform humans in social situational judgments

 

Expert ratings

Copilot

ChatGPT

Claude

Gemini

Gemini

Copilot

0.80

1

0.98

0.92

0.90

0.91

ChatGPT

0.79

 

1

0.92

0.91

0.89

Claude

0.87

  

1

0.92

0.84

Gemini

0.78

   

1

0.81

you.com

0.82

    

1

  1. The second column shows the correlation between option effectiveness rated by the chatbots and the experts. All following columns show the correlations of effectiveness ratings between the chatbots.
Back to article page